Added mcp rag, onboarding app, frontend for learning and agent inference

apps/mlstore/admin.py

@@ -1,6 +1,6 @@
 from django.contrib import admin
 from django.contrib.admin import ModelAdmin, TabularInline
-from apps.mlstore.models import AgentModel, AgentRun, Agent, AgentEvent
+from apps.mlstore.models import AgentModel, AgentRun, Agent, AgentEvent, RoleRagDocument
 
 
 class AgentInline(TabularInline):
@@ -55,3 +55,12 @@ class AgentEventAdmin(ModelAdmin):
 	search_fields = ('event_type', 'execution__uuid', 'execution__agent__model__name')
 	list_filter = ('event_type',)
 	raw_id_fields = ('execution',)
+
+
+@admin.register(RoleRagDocument)
+class RoleRagDocumentAdmin(ModelAdmin):
+	list_display = ('id', 'uuid', 'role', 'training_file', 'chunk_index', 'is_active', 'created_at')
+	search_fields = ('role__name', 'training_file__file_name')
+	list_filter = ('is_active', 'created_at')
+	raw_id_fields = ('role', 'training_file')
+	readonly_fields = ('uuid', 'created_at', 'updated_at')

apps/mlstore/consumers.py

@@ -41,6 +41,8 @@ class MLStoreConsumer(AsyncWebsocketConsumer):
                 await self.handle_fine_tune(data)
             elif action == "infer":
                 await self.handle_infer(data)
+            elif action == "onboarding_progress":
+                await self.handle_onboarding_progress(data)
             elif action in ("stop_agent", "stop"):
                 await self.handle_stop(data)
             else:
@@ -90,6 +92,16 @@ class MLStoreConsumer(AsyncWebsocketConsumer):
             return
 
         input_data = data.get("input_data") or {}
+        role_uuid = input_data.get("role_uuid")
+        if not role_uuid:
+            options = input_data.get("options") or {}
+            role_uuid = options.get("role_uuid")
+        if not role_uuid:
+            await self.send(json.dumps({
+                "type": "error",
+                "message": "role_uuid is required for inference to enable RAG"
+            }))
+            return
         execution = await self.create_run(agent, self.user, input_data)
 
         await self.send(json.dumps({
@@ -125,6 +137,35 @@ class MLStoreConsumer(AsyncWebsocketConsumer):
             "error_message": "Execution stopped by user"
         }))
 
+    async def handle_onboarding_progress(self, data):
+        execution_id = data.get("execution_id")
+        if not execution_id:
+            await self.send(json.dumps({
+                "type": "error",
+                "message": "execution_id required for onboarding_progress"
+            }))
+            return
+
+        execution = await self.get_execution(execution_id)
+        if not execution:
+            await self.send(json.dumps({
+                "type": "error",
+                "message": "Execution not found"
+            }))
+            return
+
+        content = data.get("content") or data.get("progress") or {}
+        await self.create_event(execution, "progress", content)
+        await self.channel_layer.group_send(
+            self.room_group_name,
+            {
+                "type": "mlstore_event",
+                "event_type": "progress",
+                "content": content,
+                "timestamp": timezone.now().isoformat(),
+            }
+        )
+
     async def mlstore_event(self, event):
         await self.send(json.dumps({
             "type": "mlstore_event",

apps/mlstore/migrations/0001_initial.py

@@ -2,6 +2,22 @@ import django.db.models.deletion
 import uuid
 from django.conf import settings
 from django.db import migrations, models
+from pgvector.django import VectorField
+
+
+def _create_vector_extension(apps, schema_editor):
+    if schema_editor.connection.vendor != 'postgresql':
+        return
+    with schema_editor.connection.cursor() as cursor:
+        cursor.execute('CREATE EXTENSION IF NOT EXISTS vector')
+
+
+def _drop_vector_extension(apps, schema_editor):
+    if schema_editor.connection.vendor != 'postgresql':
+        return
+    with schema_editor.connection.cursor() as cursor:
+        cursor.execute('DROP EXTENSION IF EXISTS vector')
+
 
 class Migration(migrations.Migration):
 
@@ -13,6 +29,10 @@ class Migration(migrations.Migration):
     ]
 
     operations = [
+        migrations.RunPython(
+            code=_create_vector_extension,
+            reverse_code=_drop_vector_extension,
+        ),
         migrations.CreateModel(
             name='AgentModel',
             fields=[
@@ -54,7 +74,7 @@ class Migration(migrations.Migration):
                 ('id', models.BigAutoField(primary_key=True, serialize=False)),
                 ('uuid', models.UUIDField(default=uuid.uuid4, editable=False, unique=True)),
                 ('status', models.CharField(choices=[('queued', 'Queued'), ('running', 'Running'), ('completed', 'Completed'), ('failed', 'Failed')], default='queued', max_length=20)),
-                ('input_data', models.JSONField(default=dict)),
+                ('input_data', models.JSONField(blank=True, default=dict)),
                 ('output_data', models.JSONField(blank=True, default=dict)),
                 ('error_message', models.TextField(blank=True, default='')),
                 ('started_at', models.DateTimeField(blank=True, null=True)),
@@ -83,4 +103,25 @@ class Migration(migrations.Migration):
                 'ordering': ['timestamp'],
             },
         ),
+        migrations.CreateModel(
+            name='RoleRagDocument',
+            fields=[
+                ('created_at', models.DateTimeField(auto_now_add=True, verbose_name='Created At')),
+                ('updated_at', models.DateTimeField(auto_now=True, verbose_name='Updated At')),
+                ('id', models.BigAutoField(primary_key=True, serialize=False)),
+                ('uuid', models.UUIDField(default=uuid.uuid4, editable=False, unique=True)),
+                ('content', models.TextField()),
+                ('content_hash', models.CharField(db_index=True, max_length=64)),
+                ('embedding', VectorField(blank=True, dimensions=1536, null=True)),
+                ('metadata', models.JSONField(blank=True, default=dict)),
+                ('chunk_index', models.IntegerField(default=0)),
+                ('is_active', models.BooleanField(default=True)),
+                ('role', models.ForeignKey(on_delete=django.db.models.deletion.CASCADE, related_name='rag_documents', to='orgs.role')),
+                ('training_file', models.ForeignKey(blank=True, null=True, on_delete=django.db.models.deletion.CASCADE, related_name='rag_documents', to='orgs.trainingfile')),
+            ],
+            options={
+                'verbose_name': 'Role RAG Document',
+                'verbose_name_plural': 'Role RAG Documents',
+            },
+        ),
     ]

apps/mlstore/migrations/0002_alter_agentrun_input_data.py

@@ -1,15 +0,0 @@
-from django.db import migrations, models
-
-class Migration(migrations.Migration):
-
-    dependencies = [
-        ('mlstore', '0001_initial'),
-    ]
-
-    operations = [
-        migrations.AlterField(
-            model_name='agentrun',
-            name='input_data',
-            field=models.JSONField(blank=True, default=dict),
-        ),
-    ]

apps/mlstore/models.py

@@ -1,7 +1,8 @@
-from django.db.models import BigAutoField, CASCADE, CharField, DateTimeField, ForeignKey, JSONField, Model, TextField, UUIDField
+from django.db.models import BigAutoField, BooleanField, CASCADE, CharField, DateTimeField, ForeignKey, JSONField, Model, TextField, UUIDField, IntegerField
+from pgvector.django import VectorField
 from apps.users.mixins import TimeStampMixin
 from apps.users.models import User
-from apps.orgs.models import Organization
+from apps.orgs.models import Organization, Role, TrainingFile
 from uuid import uuid4
 
 class AgentModel(Model):
@@ -100,3 +101,25 @@ class AgentEvent(Model):
 		ordering = ['timestamp']
 		verbose_name = "Agent Event"
 		verbose_name_plural = "Agent Events"
+
+
+class RoleRagDocument(TimeStampMixin, Model):
+
+    id = BigAutoField(primary_key = True)
+    uuid = UUIDField(default = uuid4, editable = False, unique = True)
+    role = ForeignKey(Role, on_delete = CASCADE, related_name = 'rag_documents')
+    training_file = ForeignKey(TrainingFile, on_delete = CASCADE, related_name = 'rag_documents', null = True, blank = True)
+
+    content = TextField()
+    content_hash = CharField(max_length = 64, db_index = True)
+    embedding = VectorField(dimensions = 1536, null = True, blank = True)
+    metadata = JSONField(default = dict, blank = True)
+    chunk_index = IntegerField(default = 0)
+    is_active = BooleanField(default = True)
+
+    class Meta:
+        verbose_name = "Role RAG Document"
+        verbose_name_plural = "Role RAG Documents"
+
+    def __str__(self) -> str:
+        return f"{self.role.name} - chunk {self.chunk_index}"

apps/mlstore/serializers.py

@@ -1,5 +1,6 @@
 from rest_framework.serializers import ModelSerializer
 from .models import AgentModel, Agent, AgentRun, AgentEvent
+from apps.orgs.serializers import OrganizationSerializer
 
 
 class AgentModelSerializer(ModelSerializer):
@@ -11,6 +12,7 @@ class AgentModelSerializer(ModelSerializer):
 
 class AgentSerializer(ModelSerializer):
     model = AgentModelSerializer(read_only=True)
+    organization = OrganizationSerializer(read_only=True)
 
     class Meta:
         model = Agent

apps/mlstore/services.py

@@ -1,19 +1,18 @@
 import asyncio
 import logging
 import os
-from typing import Any, Dict, List, Optional
+import re
+from typing import Any, Dict, List, Optional, Tuple
 from django.conf import settings
 from mcp_agent.mcp_client import MCPClient
-from .models import AgentModel
+from .models import AgentModel, RoleRagDocument
 
 logger = logging.getLogger(__name__)
 
-# Get reference to the base model cache directory
 try:
     from mcp_agent.mcp_server import BASE_MODEL_CACHE_DIR
     BASE_MODEL_CACHE = BASE_MODEL_CACHE_DIR
 except ImportError:
-    # Fallback: construct the path manually
     project_root = os.path.dirname(os.path.dirname(os.path.dirname(os.path.abspath(__file__))))
     BASE_MODEL_CACHE = os.path.join(project_root, "model", "base-model")
 
@@ -67,7 +66,6 @@ def fine_tune_model(
     except Exception as e:
         error_msg = str(e) if str(e) else f"Unknown error: {type(e).__name__}"
         logger.error(f"Fine-tune failed: {error_msg}", exc_info=True)
-        # Return a failed response instead of raising
         return {
             "status": "failed",
             "error": error_msg,
@@ -114,3 +112,294 @@ def register_model_in_db(name: str, version: str, model_path: str) -> AgentModel
     NOTE: migrations are required after the model field change prior to using this in production.
     """
     return AgentModel.objects.create(name=name, version=version, path=model_path)
+
+
+def embed_texts(texts: List[str]) -> List[List[float]]:
+    """Generate embeddings for texts using the MCP embedding service.
+    
+    Falls back to local sentence-transformers if MCP unavailable.
+    
+    Args:
+        texts: List of text strings to embed.
+    
+    Returns:
+        List of embedding vectors (list of floats).
+    
+    Raises:
+        RuntimeError: If both MCP and local embedding fail.
+    """
+    logger.info(f"Embedding {len(texts)} texts")
+    
+    try:
+        result = asyncio.run(_call_mcp("embed", {"texts": texts}))
+        embeddings = result.get("embeddings", [])
+        if embeddings and len(embeddings) == len(texts):
+            logger.info(f"Successfully embedded {len(texts)} texts via MCP")
+            return embeddings
+    except Exception as e:
+        logger.warning(f"MCP embedding failed, trying local fallback: {e}")
+    
+    try:
+        from sentence_transformers import SentenceTransformer
+        model = SentenceTransformer("all-MiniLM-L6-v2")
+        embeddings = model.encode(texts).tolist()
+        logger.info(f"Successfully embedded {len(texts)} texts via local model")
+        return embeddings
+    except Exception as e:
+        logger.error(f"Local embedding also failed: {e}")
+        raise RuntimeError(f"Failed to embed texts: {e}")
+
+
+def embed_text(text: str) -> List[float]:
+    """Generate embedding for a single text.
+    
+    Args:
+        text: Text string to embed.
+    
+    Returns:
+        Embedding vector (list of floats).
+    """
+    return embed_texts([text])[0]
+
+
+def search_similar_documents(
+    query: str,
+    role_uuid: str,
+    top_k: int = 5,
+    similarity_threshold: float = 0.0,
+) -> List[Tuple[RoleRagDocument, float]]:
+    """Search for documents similar to the query using vector similarity.
+    
+    Args:
+        query: Query text to embed and search for.
+        role_uuid: UUID of role to scope search.
+        top_k: Number of top results to return.
+        similarity_threshold: Minimum similarity score (0-1) to include results.
+    
+    Returns:
+        List of (RoleRagDocument, similarity_score) tuples, ordered by similarity DESC.
+    
+    Raises:
+        ValueError: If role not found or embedding fails.
+    """
+    from apps.orgs.models import Role
+    
+    try:
+        query_embedding = embed_text(query)
+        logger.info(f"Embedded query: '{query[:50]}...' to {len(query_embedding)}D vector")
+    except Exception as e:
+        logger.error(f"Failed to embed query: {e}")
+        raise ValueError(f"Failed to embed query: {e}")
+
+    try:
+        role = Role.objects.get(uuid=role_uuid)
+    except Role.DoesNotExist:
+        raise ValueError(f"Role with UUID {role_uuid} not found")
+
+    queryset = RoleRagDocument.objects.filter(
+        role=role,
+    )
+
+    if not queryset.exists():
+        logger.warning(f"No documents with embeddings found for role {role.uuid}")
+        return []
+
+    from django.db import connection
+
+    with connection.cursor() as cursor:
+        query_sql = """
+            SELECT id, 1 - (embedding <=> %s::vector) as similarity
+            FROM mlstore_roleragdocument
+            WHERE role_id = %s AND embedding IS NOT NULL
+            ORDER BY similarity DESC
+            LIMIT %s
+        """
+
+        cursor.execute(
+            query_sql,
+        )
+
+        doc_ids_with_scores = cursor.fetchall()
+
+    if not doc_ids_with_scores:
+        logger.info(f"No similar documents found for query in role {role.uuid}")
+        return []
+
+    filtered_docs = [
+        (doc_id, score)
+        for doc_id, score in doc_ids_with_scores
+        if score >= similarity_threshold
+    ][:top_k]
+
+    if not filtered_docs:
+        logger.info(
+            f"No documents met similarity threshold {similarity_threshold}"
+        )
+        return []
+
+    doc_ids = [doc_id for doc_id, _ in filtered_docs]
+    doc_scores = {doc_id: score for doc_id, score in filtered_docs}
+
+    documents = RoleRagDocument.objects.filter(id__in=doc_ids)
+    results = [
+        (doc, doc_scores[doc.id])
+        for doc in documents
+        if doc.id in doc_scores
+    ]
+
+    results.sort(key=lambda x: x[1], reverse=True)
+
+    logger.info(
+        f"Found {len(results)} similar documents for query "
+        f"(threshold={similarity_threshold}, top_k={top_k})"
+    )
+
+    return results
+
+
+def batch_embed_documents(
+    documents: List[RoleRagDocument],
+    batch_size: int = 32,
+    force_reembed: bool = False,
+) -> Tuple[int, int]:
+    """Batch embed documents that don't have embeddings yet.
+    
+    Args:
+        documents: List of RoleRagDocument instances to embed.
+        batch_size: Number of documents to embed per API call.
+        force_reembed: If True, re-embed documents that already have embeddings.
+    
+    Returns:
+        Tuple of (num_embedded, num_failed).
+    
+    Note:
+        Updates documents in-place with embedding values.
+    """
+    to_embed = [
+        doc for doc in documents
+        if force_reembed or not doc.embedding
+    ]
+
+    if not to_embed:
+        logger.info("No documents to embed")
+        return 0, 0
+
+    num_embedded = 0
+    num_failed = 0
+
+    for i in range(0, len(to_embed), batch_size):
+        batch = to_embed[i : i + batch_size]
+        logger.info(
+            f"Embedding batch {i // batch_size + 1} "
+            f"({len(batch)} documents)"
+        )
+
+        try:
+            texts = [doc.content for doc in batch]
+            embeddings = embed_texts(texts)
+
+            for doc, embedding in zip(batch, embeddings):
+                doc.embedding = embedding
+                num_embedded += 1
+
+            RoleRagDocument.objects.bulk_update(batch, ["embedding"], batch_size=500)
+            logger.info(f"Successfully embedded {len(batch)} documents")
+
+        except Exception as e:
+            logger.error(f"Failed to embed batch: {e}")
+            num_failed += len(batch)
+
+    logger.info(
+        f"Embedding complete: {num_embedded} embedded, {num_failed} failed"
+    )
+    return num_embedded, num_failed
+
+
+def get_context_for_query(
+    query: str,
+    role_uuid: str,
+    top_k: int = 5,
+    similarity_threshold: float = 0.5,
+) -> str:
+    """Get context string from similar documents for a query.
+    
+    Useful for augmenting prompts with retrieved context.
+    
+    Args:
+        query: Query text.
+        role_uuid: UUID of role to search within.
+        top_k: Number of top results to include.
+        similarity_threshold: Minimum similarity score.
+    
+    Returns:
+        Formatted context string with source attribution.
+    """
+    def _clean_chunk_text(text: str) -> str:
+        """Strip junk and deduplicate paragraphs to keep context lean."""
+        if not text:
+            return ""
+
+        text = re.sub(r"\[\s*Answer\s*:.*?\]", "", text, flags=re.IGNORECASE | re.DOTALL)
+
+        lines = []
+        for raw_line in text.splitlines():
+            line = raw_line.strip()
+            if not line:
+                lines.append("")
+                continue
+
+            lower = line.lower()
+
+            if line.startswith("#"):
+                continue
+            if "do you have any questions" in lower:
+                continue
+            if "feel free to ask" in lower:
+                continue
+            if "references" in lower or "sources" in lower or "wikipedia" in lower:
+                continue
+
+            lines.append(line)
+
+        cleaned = "\n".join(lines)
+
+        paragraphs = [p.strip() for p in re.split(r"\n\s*\n+", cleaned) if p.strip()]
+        seen = set()
+        unique_paragraphs: List[str] = []
+        for para in paragraphs:
+            if para in seen:
+                continue
+            seen.add(para)
+            unique_paragraphs.append(para)
+
+        return "\n\n".join(unique_paragraphs)
+    try:
+        results = search_similar_documents(
+            query=query,
+            role_uuid=role_uuid,
+            top_k=top_k,
+            similarity_threshold=similarity_threshold,
+        )
+    except Exception as e:
+        logger.warning(f"Failed to retrieve context: {e}")
+        return ""
+
+    if not results:
+        return ""
+
+    context_parts = []
+    for doc, similarity in results:
+        cleaned = _clean_chunk_text(doc.content)
+        if not cleaned:
+            continue
+
+        source = "unknown"
+        if doc.training_file:
+            source = doc.training_file.file_name
+
+        context_parts.append(
+            f"[Source: {source}, Similarity: {similarity:.2%}]\n{cleaned}\n"
+        )
+
+    context = "\n---\n".join(context_parts)
+    return context

apps/mlstore/tasks.py

@@ -1,16 +1,128 @@
 import logging
+import os
+import re
+import time
 import traceback
+from hashlib import sha256
 from asgiref.sync import async_to_sync
 from celery import shared_task
 from channels.layers import get_channel_layer
 from django.utils import timezone
+from django.db import transaction
 
-from apps.orgs.models import TrainingFile
+from apps.orgs.models import TrainingFile, Role
 from . import services
-from .models import Agent, AgentEvent, AgentModel, AgentRun
+from .models import Agent, AgentEvent, AgentModel, AgentRun, RoleRagDocument
 
 logger = logging.getLogger(__name__)
 
+
+def _get_mem_info() -> str:
+    try:
+        with open('/proc/self/status', 'r', encoding='utf-8') as f:
+            lines = f.read().splitlines()
+        mem = {line.split(':', 1)[0]: line.split(':', 1)[1].strip() for line in lines if ':' in line}
+        return f"VmRSS={mem.get('VmRSS','?')}, VmHWM={mem.get('VmHWM','?')}, VmSize={mem.get('VmSize','?')}"
+    except Exception:
+        return "mem_info_unavailable"
+
+
+def _estimate_tokens(text: str) -> int:
+    if not text:
+        return 0
+    return len(re.findall(r"\w+|[^\s\w]", text))
+
+
+def _split_semantic_units(text: str) -> list[str]:
+    paragraphs = [p.strip() for p in re.split(r"\n\s*\n+", text) if p.strip()]
+    units: list[str] = []
+    for para in paragraphs:
+        sentences = re.split(r"(?<=[.!?])\s+", para)
+        for sent in sentences:
+            sent = sent.strip()
+            if sent:
+                units.append(sent)
+    return units or paragraphs
+
+
+def _chunk_text(text: str, max_tokens: int = 400, overlap_tokens: int = 50) -> list[str]:
+    if not text:
+        return []
+
+    units = _split_semantic_units(text)
+    logger.info(
+        "Semantic chunking units=%s max_tokens=%s overlap_tokens=%s mem=%s",
+        len(units),
+        max_tokens,
+        overlap_tokens,
+        _get_mem_info(),
+    )
+
+    chunks: list[str] = []
+    current: list[str] = []
+    current_tokens = 0
+
+    for unit in units:
+        unit_tokens = _estimate_tokens(unit)
+        if unit_tokens == 0:
+            continue
+
+        if current_tokens + unit_tokens > max_tokens and current:
+            chunk = " ".join(current).strip()
+            if chunk:
+                chunks.append(chunk)
+
+            if overlap_tokens > 0:
+                overlap: list[str] = []
+                overlap_count = 0
+                for prev in reversed(current):
+                    prev_tokens = _estimate_tokens(prev)
+                    if overlap_count + prev_tokens > overlap_tokens:
+                        break
+                    overlap.insert(0, prev)
+                    overlap_count += prev_tokens
+                current = overlap
+                current_tokens = overlap_count
+            else:
+                current = []
+                current_tokens = 0
+
+        current.append(unit)
+        current_tokens += unit_tokens
+
+    if current:
+        chunk = " ".join(current).strip()
+        if chunk:
+            chunks.append(chunk)
+
+    return chunks
+
+
+def _extract_text_from_file(file_path: str, file_type: str | None) -> str:
+    file_type = (file_type or '').lower()
+    if file_type in {'txt', 'md', 'csv', 'json'}:
+        with open(file_path, 'r', encoding='utf-8', errors='ignore') as f:
+            return f.read()
+
+    if file_type == 'pdf':
+        try:
+            from PyPDF2 import PdfReader
+        except Exception as e:
+            raise RuntimeError('PyPDF2 is required to parse PDF files') from e
+        reader = PdfReader(file_path)
+        return "\n".join(page.extract_text() or "" for page in reader.pages)
+
+    if file_type in {'docx', 'doc'}:
+        try:
+            import docx
+        except Exception as e:
+            raise RuntimeError('python-docx is required to parse DOCX files') from e
+        doc = docx.Document(file_path)
+        return "\n".join(p.text for p in doc.paragraphs)
+
+    with open(file_path, 'r', encoding='utf-8', errors='ignore') as f:
+        return f.read()
+
 def _send_group_event(room_group_name: str, event_type: str, content: dict):
     channel_layer = get_channel_layer()
     async_to_sync(channel_layer.group_send)(
@@ -68,11 +180,21 @@ def start_fine_tune_run_task(execution_id: str):
     
     training_files = input_data.get("training_files") or []
     org_training_files = []
+    role_uuid = input_data.get("role_uuid")
     if not training_files and agent.organization:
-        org_training_files = list(TrainingFile.objects.filter(
+        training_files_qs = TrainingFile.objects.filter(
-            organization=agent.organization,
+            role__organization=agent.organization,
             is_processed=False
-        ).select_related('uploaded_by'))
+        ).select_related('uploaded_by', 'role')
+
+        if role_uuid:
+            try:
+                role = Role.objects.get(uuid=role_uuid, organization=agent.organization)
+                training_files_qs = training_files_qs.filter(role=role)
+            except Role.DoesNotExist:
+                logger.warning(f"Role {role_uuid} not found for organization {agent.organization.name}")
+
+        org_training_files = list(training_files_qs)
         training_files = [tf.file.path for tf in org_training_files if tf.file]
         logger.info(f"Fetched {len(training_files)} training files from organization {agent.organization.name}")
     
@@ -186,6 +308,197 @@ def start_fine_tune_run_task(execution_id: str):
         return {"status": "error", "execution_id": execution_id, "error": str(e)}
 
 
+@shared_task
+def ingest_training_file_task(training_file_uuid: str):
+    logger.info(f"Ingest task started for training_file_uuid={training_file_uuid}")
+    started_at = time.time()
+    try:
+        training_file = TrainingFile.objects.select_related('role').get(uuid=training_file_uuid)
+    except TrainingFile.DoesNotExist:
+        logger.error(f"Training file not found: {training_file_uuid}")
+        return {"status": "error", "error": "training_file_not_found"}
+
+    if training_file.is_processed:
+        logger.info(f"Training file already processed: {training_file_uuid}")
+        return {"status": "skipped", "reason": "already_processed"}
+
+    if not training_file.file:
+        logger.error(f"Training file has no file attached: {training_file_uuid}")
+        return {"status": "error", "error": "file_missing"}
+
+    try:
+        file_path = training_file.file.path
+        file_size = os.path.getsize(file_path) if os.path.exists(file_path) else 0
+        logger.info(
+            "Ingesting file: name=%s type=%s size_bytes=%s role=%s path=%s",
+            training_file.file_name,
+            training_file.file_type,
+            file_size,
+            training_file.role_id,
+            file_path,
+        )
+    except Exception as e:
+        logger.warning(f"Failed to stat training file for {training_file_uuid}: {str(e)}")
+
+    try:
+        training_file.status = 'ingesting'
+        training_file.save(update_fields=['status'])
+
+        extract_started = time.time()
+        text = _extract_text_from_file(training_file.file.path, training_file.file_type)
+        logger.info(
+            "Extracted text length=%s for training_file_uuid=%s in %.2fs mem=%s",
+            len(text),
+            training_file_uuid,
+            time.time() - extract_started,
+            _get_mem_info(),
+        )
+        chunk_started = time.time()
+        chunks = _chunk_text(text)
+        logger.info(
+            "Chunked text into %s chunks in %.2fs (sample lengths: %s) mem=%s",
+            len(chunks),
+            time.time() - chunk_started,
+            [len(c) for c in chunks[:5]],
+            _get_mem_info(),
+        )
+        if not chunks:
+            raise RuntimeError("No text extracted from file")
+
+        with transaction.atomic():
+            logger.info("Clearing existing RAG docs for training_file_uuid=%s mem=%s", training_file_uuid, _get_mem_info())
+            RoleRagDocument.objects.filter(training_file=training_file).delete()
+            logger.info("Preparing %s RAG docs for bulk_create mem=%s", len(chunks), _get_mem_info())
+            existing_hashes = set(
+                RoleRagDocument.objects.filter(role=training_file.role)
+                .values_list('content_hash', flat=True)
+            )
+            documents = []
+            skipped = 0
+            for index, chunk in enumerate(chunks):
+                content_hash = sha256(chunk.encode('utf-8')).hexdigest()
+                if content_hash in existing_hashes:
+                    skipped += 1
+                    continue
+                documents.append(
+                    RoleRagDocument(
+                        role=training_file.role,
+                        training_file=training_file,
+                        content=chunk,
+                        embedding=None,
+                        content_hash=content_hash,
+                        metadata={
+                            "file_name": training_file.file_name,
+                            "file_type": training_file.file_type,
+                            "chunk_size": len(chunk),
+                            "source": "training_file",
+                        },
+                        chunk_index=index,
+                    )
+                )
+            logger.info("Skipped %s duplicate chunks based on content_hash", skipped)
+            logger.info("Bulk creating RAG docs count=%s mem=%s", len(documents), _get_mem_info())
+            RoleRagDocument.objects.bulk_create(documents, batch_size=500)
+            training_file.status = 'chunked'
+            training_file.is_processed = True
+            training_file.save(update_fields=['status', 'is_processed'])
+
+        elapsed = time.time() - started_at
+        logger.info(
+            "Ingested training file %s into %s RAG chunks in %.2fs",
+            training_file_uuid,
+            len(chunks),
+            elapsed,
+        )
+        
+        logger.info(f"Enqueueing embedding task for training_file_uuid={training_file_uuid}")
+        embed_training_file_task.delay(training_file_uuid)
+        
+        return {"status": "completed", "chunks": len(chunks)}
+    except Exception as e:
+        elapsed = time.time() - started_at
+        logger.error(f"Failed to ingest training file {training_file_uuid}: {str(e)}", exc_info=True)
+        logger.error(f"Ingest task failed after {elapsed:.2f}s for training_file_uuid={training_file_uuid}")
+        try:
+            TrainingFile.objects.filter(uuid=training_file_uuid).update(status='failed')
+        except Exception:
+            pass
+        return {"status": "error", "error": str(e)}
+
+
+@shared_task
+def embed_training_file_task(training_file_uuid: str):
+    """Generate embeddings for all documents of a training file.
+    
+    This task is called after chunking to embed the document chunks
+    using the configured embedding provider (OpenAI, Google, or local).
+    """
+    logger.info(f"Embedding task started for training_file_uuid={training_file_uuid}")
+    started_at = time.time()
+    
+    try:
+        training_file = TrainingFile.objects.select_related('role').get(uuid=training_file_uuid)
+    except TrainingFile.DoesNotExist:
+        logger.error(f"Training file not found: {training_file_uuid}")
+        return {"status": "error", "error": "training_file_not_found"}
+    
+    try:
+        documents = list(RoleRagDocument.objects.filter(training_file=training_file))
+        
+        if not documents:
+            logger.warning(f"No RAG documents found for training_file_uuid={training_file_uuid}")
+            return {"status": "skipped", "reason": "no_documents"}
+        
+        logger.info(
+            f"Starting to embed {len(documents)} documents for training_file_uuid={training_file_uuid} "
+            f"mem={_get_mem_info()}"
+        )
+        
+        num_embedded, num_failed = services.batch_embed_documents(documents, batch_size=32)
+        
+        if num_failed == 0:
+            training_file.status = 'embedded'
+            training_file.save(update_fields=['status'])
+            logger.info(f"Successfully embedded all documents for training_file_uuid={training_file_uuid}")
+        elif num_embedded > 0:
+            training_file.status = 'embedded'
+            training_file.save(update_fields=['status'])
+            logger.warning(
+                f"Partially embedded {num_embedded} documents, {num_failed} failed "
+                f"for training_file_uuid={training_file_uuid}"
+            )
+        else:
+            training_file.status = 'failed'
+            training_file.save(update_fields=['status'])
+            logger.error(f"Failed to embed any documents for training_file_uuid={training_file_uuid}")
+            return {"status": "error", "error": "embedding_failed", "num_failed": num_failed}
+        
+        elapsed = time.time() - started_at
+        logger.info(
+            f"Embedding task completed for {training_file_uuid}: "
+            f"embedded={num_embedded}, failed={num_failed}, time={elapsed:.2f}s"
+        )
+        
+        return {
+            "status": "completed",
+            "num_embedded": num_embedded,
+            "num_failed": num_failed,
+            "elapsed": elapsed,
+        }
+        
+    except Exception as e:
+        elapsed = time.time() - started_at
+        logger.error(
+            f"Failed to embed training file {training_file_uuid}: {str(e)}",
+            exc_info=True
+        )
+        try:
+            TrainingFile.objects.filter(uuid=training_file_uuid).update(status='failed')
+        except Exception:
+            pass
+        return {"status": "error", "error": str(e), "elapsed": elapsed}
+
+
 @shared_task
 def infer_run_task(execution_id: str):
     logger.info(f"Inference run task started for execution: {execution_id}")
@@ -207,9 +520,70 @@ def infer_run_task(execution_id: str):
 
     input_data = execution.input_data or {}
     prompt = input_data.get("prompt") or input_data.get("query") or ""
-    options = input_data.get("options") or {}
+    options = dict(input_data.get("options") or {})
+    role_uuid = input_data.get("role_uuid") or options.get("role_uuid")
+    rag_top_k = int(input_data.get("rag_top_k", 5))
+    rag_similarity_threshold = float(input_data.get("rag_similarity_threshold", 0.5))
+
+    options.setdefault("temperature", 0.2)
+    options.setdefault("top_p", 0.9)
+    options.setdefault("max_tokens", 200)
+    options.setdefault("stop", ["\n\n", "References:", "Sources:"])
+
     logger.info(f"Prompt length: {len(prompt)} characters")
 
+    if not role_uuid:
+        logger.warning(f"No role_uuid provided for inference run {execution_id}")
+        execution.status = "failed"
+        execution.error_message = "role_uuid_required"
+        execution.completed_at = timezone.now()
+        execution.save()
+        _update_agent_status(agent, "failed")
+        _send_group_event(room_group_name, "error", {"execution_id": str(execution.uuid), "error": "role_uuid_required"})
+        _persist_event(execution, "error", {"execution_id": str(execution.uuid), "error": "role_uuid_required"})
+        async_to_sync(get_channel_layer().group_send)(
+            room_group_name,
+            {
+                "type": "mlstore_error",
+                "execution_id": str(execution.uuid),
+                "error_message": "role_uuid_required",
+            },
+        )
+        return {"status": "failed", "execution_id": execution_id, "error": "role_uuid_required"}
+
+    if role_uuid and prompt:
+        try:
+            context = services.get_context_for_query(
+                query=prompt,
+                role_uuid=str(role_uuid),
+                top_k=rag_top_k,
+                similarity_threshold=rag_similarity_threshold,
+            )
+            if context:
+                logger.info(f"RAG context retrieved for role={role_uuid} (top_k={rag_top_k})")
+                prompt = (
+                    "You are a technical assistant.\n\n"
+                    "Answer the question using ONLY the information in the context.\n"
+                    "Do NOT:\n"
+                    "- ask follow-up questions\n"
+                    "- include hashtags\n"
+                    "- include references or sources\n"
+                    "- repeat the question\n"
+                    "- add headings or sections\n"
+                    "- add information not present in the context\n\n"
+                    "Answer in 3-6 concise sentences.\n"
+                    "If the context is insufficient, say: \"The context does not provide enough information.\"\n\n"
+                    "Context:\n"
+                    f"{context}\n\n"
+                    "Question:\n"
+                    f"{prompt}\n\n"
+                    "Answer:"
+                )
+            else:
+                logger.info(f"No RAG context found for role={role_uuid}")
+        except Exception as e:
+            logger.warning(f"RAG context retrieval failed for role={role_uuid}: {e}")
+
     if not prompt:
         logger.warning(f"No prompt provided for inference run {execution_id}")
         execution.status = "failed"

apps/mlstore/tests/test_api.py

@@ -0,0 +1,91 @@
+from unittest.mock import patch
+from django.contrib.auth import get_user_model
+from django.test import TestCase
+from rest_framework.test import APIRequestFactory, force_authenticate
+from rest_framework.status import HTTP_200_OK, HTTP_400_BAD_REQUEST, HTTP_403_FORBIDDEN, HTTP_404_NOT_FOUND
+
+from apps.orgs.models import Organization, Role
+from apps.mlstore.models import AgentModel, Agent, AgentRun, AgentEvent, RoleRagDocument
+from apps.mlstore.viewsets import AgentViewSet, AgentRunViewSet
+
+User = get_user_model()
+
+
+class MLStoreAPITests(TestCase):
+    def setUp(self):
+        self.factory = APIRequestFactory()
+        self.user = User.objects.create_user(email_address='user@example.com', password='pass')
+        self.other = User.objects.create_user(email_address='other@example.com', password='pass')
+        self.manager = User.objects.create_user(email_address='manager@example.com', password='pass', is_manager=True)
+        self.org = Organization.objects.create(name='Org', owner=self.manager)
+        self.role = Role.objects.create(name='Engineer', organization=self.org)
+        self.model = AgentModel.objects.create(name='test-model', version='v1', path='model.gguf')
+        self.agent = Agent.objects.create(model=self.model, organization=self.org)
+
+    def test_agents_list_requires_auth(self):
+        view = AgentViewSet.as_view({'get': 'list'})
+        request = self.factory.get('/')
+        response = view(request)
+        self.assertEqual(response.status_code, HTTP_403_FORBIDDEN)
+
+    def test_agents_list_authenticated(self):
+        view = AgentViewSet.as_view({'get': 'list'})
+        request = self.factory.get('/')
+        force_authenticate(request, user=self.user)
+        response = view(request)
+        self.assertEqual(response.status_code, HTTP_200_OK)
+
+    def test_agent_runs_scoped_to_user(self):
+        AgentRun.objects.create(agent=self.agent, user=self.user)
+        AgentRun.objects.create(agent=self.agent, user=self.other)
+        view = AgentRunViewSet.as_view({'get': 'list'})
+        request = self.factory.get('/')
+        force_authenticate(request, user=self.user)
+        response = view(request)
+        self.assertEqual(response.status_code, HTTP_200_OK)
+        self.assertEqual(len(response.data), 1)
+
+    def test_agent_run_events(self):
+        run = AgentRun.objects.create(agent=self.agent, user=self.user)
+        AgentEvent.objects.create(execution=run, event_type='message', content={'msg': 'hi'})
+        view = AgentRunViewSet.as_view({'get': 'events'})
+        request = self.factory.get('/')
+        force_authenticate(request, user=self.user)
+        response = view(request, uuid=str(run.uuid))
+        self.assertEqual(response.status_code, HTTP_200_OK)
+        self.assertEqual(len(response.data), 1)
+
+    def test_retrieve_context_missing_params(self):
+        view = AgentRunViewSet.as_view({'post': 'retrieve_context'})
+        request = self.factory.post('/', {})
+        force_authenticate(request, user=self.user)
+        response = view(request)
+        self.assertEqual(response.status_code, HTTP_400_BAD_REQUEST)
+
+    def test_retrieve_context_role_not_found(self):
+        view = AgentRunViewSet.as_view({'post': 'retrieve_context'})
+        request = self.factory.post('/', {'query': 'q', 'role_uuid': '00000000-0000-0000-0000-000000000000'})
+        force_authenticate(request, user=self.user)
+        response = view(request)
+        self.assertEqual(response.status_code, HTTP_404_NOT_FOUND)
+
+    @patch('apps.mlstore.viewsets.services.search_similar_documents')
+    @patch('apps.mlstore.viewsets.services.get_context_for_query')
+    def test_retrieve_context_success(self, mock_context, mock_search):
+        doc = RoleRagDocument.objects.create(
+            role=self.role,
+            content='chunk',
+            content_hash='hash',
+            chunk_index=0,
+        )
+        mock_search.return_value = [(doc, 0.9)]
+        mock_context.return_value = 'context text'
+
+        view = AgentRunViewSet.as_view({'post': 'retrieve_context'})
+        payload = {'query': 'hello', 'role_uuid': str(self.role.uuid)}
+        request = self.factory.post('/', payload, format='json')
+        force_authenticate(request, user=self.user)
+        response = view(request)
+        self.assertEqual(response.status_code, HTTP_200_OK)
+        self.assertEqual(response.data.get('num_results'), 1)
+        self.assertEqual(response.data.get('context'), 'context text')

apps/mlstore/tests/test_models.py

@@ -0,0 +1,41 @@
+from django.test import TestCase
+from django.contrib.auth import get_user_model
+from apps.orgs.models import Organization, Role
+from apps.mlstore.models import AgentModel, Agent, AgentRun, AgentEvent, RoleRagDocument
+
+User = get_user_model()
+
+
+class MLStoreModelTests(TestCase):
+    def setUp(self):
+        self.user = User.objects.create_user(email_address='user@example.com', password='pass')
+        self.manager = User.objects.create_user(email_address='manager@example.com', password='pass', is_manager=True)
+        self.org = Organization.objects.create(name='Org', owner=self.manager)
+        self.role = Role.objects.create(name='Engineer', organization=self.org)
+        self.model = AgentModel.objects.create(name='test-model', version='v1', path='model.gguf')
+        self.agent = Agent.objects.create(model=self.model, organization=self.org)
+
+    def test_agent_model_str(self):
+        self.assertEqual(str(self.model), 'test-model')
+
+    def test_agent_str(self):
+        self.assertIn(self.model.name, str(self.agent))
+
+    def test_agent_run_str(self):
+        run = AgentRun.objects.create(agent=self.agent, user=self.user)
+        self.assertIn(str(run.uuid), str(run))
+        self.assertIn(str(self.agent), str(run))
+
+    def test_agent_event_str(self):
+        run = AgentRun.objects.create(agent=self.agent, user=self.user)
+        evt = AgentEvent.objects.create(execution=run, event_type='message', content={'msg': 'hi'})
+        self.assertIn('message', str(evt))
+
+    def test_role_rag_document_str(self):
+        doc = RoleRagDocument.objects.create(
+            role=self.role,
+            content='chunk',
+            content_hash='hash',
+            chunk_index=0,
+        )
+        self.assertIn(self.role.name, str(doc))

apps/mlstore/viewsets.py

@@ -1,9 +1,13 @@
 from rest_framework.viewsets import ModelViewSet
 from rest_framework.permissions import IsAuthenticated
-from .models import Agent, AgentRun, AgentEvent
-from .serializers import AgentSerializer, AgentRunSerializer, AgentEventSerializer
 from rest_framework.decorators import action
 from rest_framework.response import Response
+from rest_framework import status
+from .models import Agent, AgentRun, AgentEvent
+from .serializers import AgentSerializer, AgentRunSerializer, AgentEventSerializer
+from . import services
+from apps.orgs.models import Role
+
 
 class AgentViewSet(ModelViewSet):
     queryset = Agent.objects.all()
@@ -27,3 +31,109 @@ class AgentRunViewSet(ModelViewSet):
         events = AgentEvent.objects.filter(execution=run).order_by('timestamp')
         serializer = AgentEventSerializer(events, many=True)
         return Response(serializer.data)
+
+    @action(detail=False, methods=['post'], url_path='retrieve-context')
+    def retrieve_context(self, request):
+        """Retrieve context documents from RAG using semantic similarity.
+        
+        Request body:
+        {
+            "query": "search query text",
+            "role_uuid": "role-uuid",
+            "top_k": 5,  # optional, default 5
+            "similarity_threshold": 0.5  # optional, default 0.5
+        }
+        
+        Returns:
+        {
+            "query": "search query text",
+            "context": "formatted context string with sources",
+            "documents": [
+                {
+                    "id": 123,
+                    "content": "chunk text",
+                    "similarity": 0.87,
+                    "source": "filename.pdf",
+                    "chunk_index": 0
+                },
+                ...
+            ]
+        }
+        """
+        query = request.data.get('query', '').strip()
+        role_uuid = request.data.get('role_uuid', '').strip()
+        top_k = request.data.get('top_k', 5)
+        similarity_threshold = request.data.get('similarity_threshold', 0.5)
+        
+        if not query:
+            return Response(
+                {"error": "query is required"},
+                status=status.HTTP_400_BAD_REQUEST
+            )
+        
+        if not role_uuid:
+            return Response(
+                {"error": "role_uuid is required"},
+                status=status.HTTP_400_BAD_REQUEST
+            )
+        
+        try:
+            # Validate role exists and user has access
+            role = Role.objects.get(uuid=role_uuid)
+            # You can add additional permission checks here if needed
+            
+            # Search for similar documents
+            results = services.search_similar_documents(
+                query=query,
+                role_uuid=role_uuid,
+                top_k=top_k,
+                similarity_threshold=similarity_threshold
+            )
+            
+            # Format response
+            documents = []
+            for doc, similarity in results:
+                documents.append({
+                    "id": doc.id,
+                    "uuid": str(doc.uuid),
+                    "content": doc.content,
+                    "similarity": float(similarity),
+                    "source": doc.training_file.file_name if doc.training_file else "unknown",
+                    "chunk_index": doc.chunk_index,
+                    "metadata": doc.metadata,
+                })
+            
+            # Get formatted context string
+            context = services.get_context_for_query(
+                query=query,
+                role_uuid=role_uuid,
+                top_k=top_k,
+                similarity_threshold=similarity_threshold
+            )
+            
+            return Response({
+                "query": query,
+                "role_uuid": role_uuid,
+                "num_results": len(documents),
+                "context": context,
+                "documents": documents,
+            })
+            
+        except Role.DoesNotExist:
+            return Response(
+                {"error": f"Role with UUID {role_uuid} not found"},
+                status=status.HTTP_404_NOT_FOUND
+            )
+        except ValueError as e:
+            return Response(
+                {"error": str(e)},
+                status=status.HTTP_400_BAD_REQUEST
+            )
+        except Exception as e:
+            import logging
+            logging.exception("Error retrieving context")
+            return Response(
+                {"error": "Failed to retrieve context", "detail": str(e)},
+                status=status.HTTP_500_INTERNAL_SERVER_ERROR
+            )
+

apps/onboarding/admin.py

@@ -0,0 +1,47 @@
+from django.contrib import admin
+from django.contrib.admin import ModelAdmin, TabularInline
+from .models import OnboardingFlow, OnboardingPage, OnboardingField, OnboardingSession
+
+
+class OnboardingPageInline(TabularInline):
+    model = OnboardingPage
+    extra = 0
+
+
+class OnboardingFieldInline(TabularInline):
+    model = OnboardingField
+    extra = 0
+
+
+@admin.register(OnboardingFlow)
+class OnboardingFlowAdmin(ModelAdmin):
+    list_display = ('id', 'uuid', 'title', 'role', 'status')
+    search_fields = ('title', 'role__name')
+    list_filter = ('status',)
+    inlines = (OnboardingPageInline,)
+    readonly_fields = ('uuid',)
+
+
+@admin.register(OnboardingPage)
+class OnboardingPageAdmin(ModelAdmin):
+    list_display = ('id', 'uuid', 'title', 'flow', 'order')
+    search_fields = ('title', 'flow__title')
+    list_filter = ('flow',)
+    inlines = (OnboardingFieldInline,)
+    readonly_fields = ('uuid',)
+
+
+@admin.register(OnboardingField)
+class OnboardingFieldAdmin(ModelAdmin):
+    list_display = ('id', 'uuid', 'label', 'page', 'field_type', 'required')
+    search_fields = ('label', 'page__title')
+    list_filter = ('field_type',)
+    readonly_fields = ('uuid',)
+
+
+@admin.register(OnboardingSession)
+class OnboardingSessionAdmin(ModelAdmin):
+    list_display = ('id', 'uuid', 'flow', 'user', 'status', 'current_page_order')
+    search_fields = ('flow__title', 'user__email_address')
+    list_filter = ('status',)
+    readonly_fields = ('uuid',)

apps/onboarding/apps.py

@@ -0,0 +1,6 @@
+from django.apps import AppConfig
+
+
+class OnboardingConfig(AppConfig):
+    default_auto_field = 'django.db.models.BigAutoField'
+    name = 'apps.onboarding'

apps/onboarding/migrations/0001_initial.py

@@ -0,0 +1,100 @@
+from django.db import migrations, models
+import django.db.models.deletion
+import uuid
+
+class Migration(migrations.Migration):
+    initial = True
+
+    dependencies = [
+        ('orgs', '0001_initial'),
+        ('mlstore', '0001_initial'),
+        ('users', '0001_initial'),
+    ]
+
+    operations = [
+        migrations.CreateModel(
+            name='OnboardingFlow',
+            fields=[
+                ('id', models.BigAutoField(primary_key=True, serialize=False)),
+                ('created_at', models.DateTimeField(auto_now_add=True, verbose_name='Created At')),
+                ('updated_at', models.DateTimeField(auto_now=True, verbose_name='Updated At')),
+                ('uuid', models.UUIDField(default=uuid.uuid4, editable=False, unique=True)),
+                ('title', models.CharField(max_length=255)),
+                ('description', models.TextField(blank=True, default='')),
+                ('status', models.CharField(choices=[('draft', 'Draft'), ('published', 'Published'), ('archived', 'Archived')], default='draft', max_length=20)),
+                ('agent', models.ForeignKey(blank=True, null=True, on_delete=django.db.models.deletion.CASCADE, related_name='onboarding_flows', to='mlstore.agent')),
+                ('role', models.ForeignKey(on_delete=django.db.models.deletion.CASCADE, related_name='onboarding_flows', to='orgs.role')),
+            ],
+            options={
+                'verbose_name': 'Onboarding Flow',
+                'verbose_name_plural': 'Onboarding Flows',
+                'ordering': ['-created_at'],
+            },
+        ),
+        migrations.CreateModel(
+            name='OnboardingPage',
+            fields=[
+                ('id', models.BigAutoField(primary_key=True, serialize=False)),
+                ('created_at', models.DateTimeField(auto_now_add=True, verbose_name='Created At')),
+                ('updated_at', models.DateTimeField(auto_now=True, verbose_name='Updated At')),
+                ('uuid', models.UUIDField(default=uuid.uuid4, editable=False, unique=True)),
+                ('order', models.IntegerField(default=0)),
+                ('title', models.CharField(max_length=255)),
+                ('body', models.TextField(blank=True, default='')),
+                ('meta', models.JSONField(blank=True, default=dict)),
+                ('flow', models.ForeignKey(on_delete=django.db.models.deletion.CASCADE, related_name='pages', to='onboarding.onboardingflow')),
+            ],
+            options={
+                'verbose_name': 'Onboarding Page',
+                'verbose_name_plural': 'Onboarding Pages',
+                'ordering': ['order', 'created_at'],
+            },
+        ),
+        migrations.CreateModel(
+            name='OnboardingField',
+            fields=[
+                ('id', models.BigAutoField(primary_key=True, serialize=False)),
+                ('created_at', models.DateTimeField(auto_now_add=True, verbose_name='Created At')),
+                ('updated_at', models.DateTimeField(auto_now=True, verbose_name='Updated At')),
+                ('uuid', models.UUIDField(default=uuid.uuid4, editable=False, unique=True)),
+                ('order', models.IntegerField(default=0)),
+                ('key', models.CharField(max_length=120)),
+                ('label', models.CharField(max_length=255)),
+                ('field_type', models.CharField(choices=[('text', 'Text'), ('textarea', 'Textarea'), ('select', 'Select'), ('multiselect', 'Multi Select'), ('number', 'Number'), ('boolean', 'Boolean'), ('date', 'Date')], default='text', max_length=30)),
+                ('required', models.BooleanField(default=False)),
+                ('help_text', models.TextField(blank=True, default='')),
+                ('placeholder', models.CharField(blank=True, default='', max_length=255)),
+                ('options', models.JSONField(blank=True, default=list)),
+                ('default_value', models.JSONField(blank=True, null=True, default=None)),
+                ('validation', models.JSONField(blank=True, default=dict)),
+                ('page', models.ForeignKey(on_delete=django.db.models.deletion.CASCADE, related_name='fields', to='onboarding.onboardingpage')),
+            ],
+            options={
+                'verbose_name': 'Onboarding Field',
+                'verbose_name_plural': 'Onboarding Fields',
+                'ordering': ['order', 'created_at'],
+                'unique_together': {('page', 'key')},
+            },
+        ),
+        migrations.CreateModel(
+            name='OnboardingSession',
+            fields=[
+                ('id', models.BigAutoField(primary_key=True, serialize=False)),
+                ('created_at', models.DateTimeField(auto_now_add=True, verbose_name='Created At')),
+                ('updated_at', models.DateTimeField(auto_now=True, verbose_name='Updated At')),
+                ('uuid', models.UUIDField(default=uuid.uuid4, editable=False, unique=True)),
+                ('status', models.CharField(choices=[('in_progress', 'In Progress'), ('completed', 'Completed'), ('abandoned', 'Abandoned')], default='in_progress', max_length=20)),
+                ('current_page_order', models.IntegerField(default=0)),
+                ('responses', models.JSONField(blank=True, default=dict)),
+                ('completed_at', models.DateTimeField(blank=True, null=True)),
+                ('agent_run', models.ForeignKey(blank=True, null=True, on_delete=django.db.models.deletion.CASCADE, related_name='onboarding_sessions', to='mlstore.agentrun')),
+                ('flow', models.ForeignKey(on_delete=django.db.models.deletion.CASCADE, related_name='sessions', to='onboarding.onboardingflow')),
+                ('user', models.ForeignKey(on_delete=django.db.models.deletion.CASCADE, related_name='onboarding_sessions', to='users.user')),
+            ],
+            options={
+                'verbose_name': 'Onboarding Session',
+                'verbose_name_plural': 'Onboarding Sessions',
+                'ordering': ['-created_at'],
+            },
+        ),
+    ]

apps/onboarding/models.py

@@ -0,0 +1,121 @@
+from uuid import uuid4
+from django.db.models import (
+    BigAutoField,
+    BooleanField,
+    CASCADE,
+    CharField,
+    DateTimeField,
+    ForeignKey,
+    IntegerField,
+    JSONField,
+    Model,
+    TextField,
+    UUIDField,
+)
+from apps.users.mixins import TimeStampMixin
+from apps.users.models import User
+from apps.orgs.models import Role
+from apps.mlstore.models import Agent, AgentRun
+
+
+class OnboardingFlow(TimeStampMixin, Model):
+    STATUS_CHOICES = [
+        ('draft', 'Draft'),
+        ('published', 'Published'),
+        ('archived', 'Archived'),
+    ]
+
+    id = BigAutoField(primary_key=True)
+    uuid = UUIDField(default=uuid4, editable=False, unique=True)
+    role = ForeignKey(Role, on_delete=CASCADE, related_name='onboarding_flows')
+    agent = ForeignKey(Agent, on_delete=CASCADE, related_name='onboarding_flows', null=True, blank=True)
+    title = CharField(max_length=255)
+    description = TextField(blank=True, default='')
+    status = CharField(max_length=20, choices=STATUS_CHOICES, default='draft')
+
+    class Meta:
+        verbose_name = 'Onboarding Flow'
+        verbose_name_plural = 'Onboarding Flows'
+        ordering = ['-created_at']
+
+    def __str__(self) -> str:
+        return f'{self.title} ({self.role.name})'
+
+
+class OnboardingPage(TimeStampMixin, Model):
+    id = BigAutoField(primary_key=True)
+    uuid = UUIDField(default=uuid4, editable=False, unique=True)
+    flow = ForeignKey(OnboardingFlow, on_delete=CASCADE, related_name='pages')
+    order = IntegerField(default=0)
+    title = CharField(max_length=255)
+    body = TextField(blank=True, default='')
+    meta = JSONField(default=dict, blank=True)
+
+    class Meta:
+        verbose_name = 'Onboarding Page'
+        verbose_name_plural = 'Onboarding Pages'
+        ordering = ['order', 'created_at']
+
+    def __str__(self) -> str:
+        return f'{self.flow.title} - {self.title}'
+
+
+class OnboardingField(TimeStampMixin, Model):
+    FIELD_TYPES = [
+        ('text', 'Text'),
+        ('textarea', 'Textarea'),
+        ('select', 'Select'),
+        ('multiselect', 'Multi Select'),
+        ('number', 'Number'),
+        ('boolean', 'Boolean'),
+        ('date', 'Date'),
+    ]
+
+    id = BigAutoField(primary_key=True)
+    uuid = UUIDField(default=uuid4, editable=False, unique=True)
+    page = ForeignKey(OnboardingPage, on_delete=CASCADE, related_name='fields')
+    order = IntegerField(default=0)
+    key = CharField(max_length=120)
+    label = CharField(max_length=255)
+    field_type = CharField(max_length=30, choices=FIELD_TYPES, default='text')
+    required = BooleanField(default=False)
+    help_text = TextField(blank=True, default='')
+    placeholder = CharField(max_length=255, blank=True, default='')
+    options = JSONField(default=list, blank=True)
+    default_value = JSONField(null=True, blank=True, default=None)
+    validation = JSONField(default=dict, blank=True)
+
+    class Meta:
+        verbose_name = 'Onboarding Field'
+        verbose_name_plural = 'Onboarding Fields'
+        ordering = ['order', 'created_at']
+        unique_together = [['page', 'key']]
+
+    def __str__(self) -> str:
+        return f'{self.page.title} - {self.label}'
+
+
+class OnboardingSession(TimeStampMixin, Model):
+    STATUS_CHOICES = [
+        ('in_progress', 'In Progress'),
+        ('completed', 'Completed'),
+        ('abandoned', 'Abandoned'),
+    ]
+
+    id = BigAutoField(primary_key=True)
+    uuid = UUIDField(default=uuid4, editable=False, unique=True)
+    flow = ForeignKey(OnboardingFlow, on_delete=CASCADE, related_name='sessions')
+    user = ForeignKey(User, on_delete=CASCADE, related_name='onboarding_sessions')
+    agent_run = ForeignKey(AgentRun, on_delete=CASCADE, related_name='onboarding_sessions', null=True, blank=True)
+    status = CharField(max_length=20, choices=STATUS_CHOICES, default='in_progress')
+    current_page_order = IntegerField(default=0)
+    responses = JSONField(default=dict, blank=True)
+    completed_at = DateTimeField(null=True, blank=True)
+
+    class Meta:
+        verbose_name = 'Onboarding Session'
+        verbose_name_plural = 'Onboarding Sessions'
+        ordering = ['-created_at']
+
+    def __str__(self) -> str:
+        return f'{self.user.email_address} - {self.flow.title}'

apps/onboarding/serializers.py

@@ -0,0 +1,105 @@
+from rest_framework import serializers
+from .models import OnboardingFlow, OnboardingPage, OnboardingField, OnboardingSession
+from apps.orgs.models import Role
+from apps.mlstore.models import Agent
+
+
+class OnboardingFieldSerializer(serializers.ModelSerializer):
+    page = serializers.SlugRelatedField(slug_field='uuid', queryset=OnboardingPage.objects.all())
+    class Meta:
+        model = OnboardingField
+        fields = [
+            'id',
+            'uuid',
+            'page',
+            'order',
+            'key',
+            'label',
+            'field_type',
+            'required',
+            'help_text',
+            'placeholder',
+            'options',
+            'default_value',
+            'validation',
+        ]
+        read_only_fields = ['id', 'uuid']
+
+
+class OnboardingPageSerializer(serializers.ModelSerializer):
+    fields = OnboardingFieldSerializer(many=True, read_only=True)
+    flow = serializers.SlugRelatedField(slug_field='uuid', queryset=OnboardingFlow.objects.all())
+
+    class Meta:
+        model = OnboardingPage
+        fields = [
+            'id',
+            'uuid',
+            'flow',
+            'order',
+            'title',
+            'body',
+            'meta',
+            'fields',
+        ]
+        read_only_fields = ['id', 'uuid']
+
+
+class OnboardingFlowSerializer(serializers.ModelSerializer):
+    role = serializers.SlugRelatedField(slug_field='uuid', queryset=Role.objects.all())
+    agent = serializers.SlugRelatedField(slug_field='uuid', queryset=Agent.objects.all(), allow_null=True, required=False)
+    class Meta:
+        model = OnboardingFlow
+        fields = [
+            'id',
+            'uuid',
+            'role',
+            'agent',
+            'title',
+            'description',
+            'status',
+            'created_at',
+            'updated_at',
+        ]
+        read_only_fields = ['id', 'uuid', 'created_at', 'updated_at']
+
+
+class OnboardingFlowDetailSerializer(OnboardingFlowSerializer):
+    pages = OnboardingPageSerializer(many=True, read_only=True)
+
+    class Meta(OnboardingFlowSerializer.Meta):
+        fields = OnboardingFlowSerializer.Meta.fields + ['pages']
+
+
+class OnboardingSessionSerializer(serializers.ModelSerializer):
+    flow = serializers.SlugRelatedField(slug_field='uuid', queryset=OnboardingFlow.objects.all())
+    user = serializers.SlugRelatedField(slug_field='uuid', read_only=True)
+    agent_run = serializers.SlugRelatedField(slug_field='uuid', read_only=True)
+    class Meta:
+        model = OnboardingSession
+        fields = [
+            'id',
+            'uuid',
+            'flow',
+            'user',
+            'agent_run',
+            'status',
+            'current_page_order',
+            'responses',
+            'created_at',
+            'updated_at',
+            'completed_at',
+        ]
+        read_only_fields = ['id', 'uuid', 'user', 'agent_run', 'created_at', 'updated_at', 'completed_at']
+
+
+class OnboardingSubmissionSerializer(serializers.Serializer):
+    page_uuid = serializers.CharField()
+    responses = serializers.DictField()
+    mark_complete = serializers.BooleanField(required=False, default=False)
+
+
+class OnboardingFeedbackSerializer(serializers.Serializer):
+    page_uuid = serializers.CharField()
+    responses = serializers.DictField()
+    question = serializers.CharField(required=False, allow_blank=True, default='')

apps/onboarding/tests/test_api.py

@@ -0,0 +1,124 @@
+from django.contrib.auth import get_user_model
+from django.test import TestCase
+from rest_framework.test import APIRequestFactory, force_authenticate
+from rest_framework.status import HTTP_200_OK, HTTP_201_CREATED, HTTP_403_FORBIDDEN
+
+from apps.orgs.models import Organization, Role
+from apps.mlstore.models import AgentModel, Agent
+from apps.onboarding.models import OnboardingFlow, OnboardingPage, OnboardingSession
+from apps.onboarding.viewsets import OnboardingFlowViewSet, OnboardingSessionViewSet
+
+User = get_user_model()
+
+
+class OnboardingAPITests(TestCase):
+    def setUp(self):
+        self.factory = APIRequestFactory()
+        self.user = User.objects.create_user(email_address='user@example.com', password='pass')
+        self.manager = User.objects.create_user(email_address='manager@example.com', password='pass', is_manager=True)
+        self.org = Organization.objects.create(name='Org', owner=self.manager)
+        self.role = Role.objects.create(name='Engineer', organization=self.org)
+        self.model = AgentModel.objects.create(name='test-model', version='v1', path='model.gguf')
+        self.agent = Agent.objects.create(model=self.model, organization=self.org)
+
+    def test_create_flow(self):
+        view = OnboardingFlowViewSet.as_view({'post': 'create'})
+        data = {
+            'role': str(self.role.uuid),
+            'agent': str(self.agent.uuid),
+            'title': 'Flow',
+            'description': 'Desc',
+            'status': 'draft',
+        }
+        request = self.factory.post('/', data)
+        force_authenticate(request, user=self.manager)
+        response = view(request)
+        self.assertIn(response.status_code, (HTTP_200_OK, HTTP_201_CREATED))
+        self.assertTrue(OnboardingFlow.objects.filter(title='Flow').exists())
+
+    def test_pages_action(self):
+        flow = OnboardingFlow.objects.create(role=self.role, agent=self.agent, title='Flow', description='')
+        OnboardingPage.objects.create(flow=flow, order=0, title='Page 1', body='Body')
+        view = OnboardingFlowViewSet.as_view({'get': 'pages'})
+        request = self.factory.get('/')
+        force_authenticate(request, user=self.manager)
+        response = view(request, uuid=str(flow.uuid))
+        self.assertEqual(response.status_code, HTTP_200_OK)
+        self.assertEqual(len(response.data.get('pages', [])), 1)
+
+
+
+    def test_create_session(self):
+        flow = OnboardingFlow.objects.create(role=self.role, agent=self.agent, title='Flow', description='')
+        view = OnboardingSessionViewSet.as_view({'post': 'create'})
+        request = self.factory.post('/', {'flow': str(flow.uuid)})
+        force_authenticate(request, user=self.user)
+        response = view(request)
+        self.assertIn(response.status_code, (HTTP_200_OK, HTTP_201_CREATED))
+        self.assertTrue(OnboardingSession.objects.filter(flow=flow, user=self.user).exists())
+
+    def test_submit_updates_session(self):
+        flow = OnboardingFlow.objects.create(role=self.role, agent=self.agent, title='Flow', description='')
+        page = OnboardingPage.objects.create(flow=flow, order=0, title='Page 1', body='Body')
+        session = OnboardingSession.objects.create(flow=flow, user=self.user)
+
+        view = OnboardingSessionViewSet.as_view({'post': 'submit'})
+        payload = {'page_uuid': str(page.uuid), 'responses': {'q1': 'a'}, 'mark_complete': True}
+        request = self.factory.post('/', payload, format='json')
+        force_authenticate(request, user=self.user)
+        response = view(request, uuid=str(session.uuid))
+        self.assertEqual(response.status_code, HTTP_200_OK)
+        session.refresh_from_db()
+        self.assertEqual(session.status, 'completed')
+        self.assertIn(str(page.uuid), session.responses)
+
+    def test_publish_flow_as_manager(self):
+        flow = OnboardingFlow.objects.create(role=self.role, agent=self.agent, title='Flow', description='')
+        self.assertEqual(flow.status, 'draft')
+        view = OnboardingFlowViewSet.as_view({'post': 'publish'})
+        request = self.factory.post('/')
+        force_authenticate(request, user=self.manager)
+        response = view(request, uuid=str(flow.uuid))
+        self.assertEqual(response.status_code, HTTP_200_OK)
+        flow.refresh_from_db()
+        self.assertEqual(flow.status, 'published')
+
+    def test_publish_flow_requires_manager(self):
+        flow = OnboardingFlow.objects.create(role=self.role, agent=self.agent, title='Flow', description='')
+        view = OnboardingFlowViewSet.as_view({'post': 'publish'})
+        request = self.factory.post('/')
+        force_authenticate(request, user=self.user)
+        response = view(request, uuid=str(flow.uuid))
+        self.assertEqual(response.status_code, HTTP_403_FORBIDDEN)
+
+    def test_get_or_create_session_creates_when_missing(self):
+        flow = OnboardingFlow.objects.create(role=self.role, agent=self.agent, title='Flow', description='')
+        view = OnboardingSessionViewSet.as_view({'post': 'get_or_create'})
+        request = self.factory.post('/', {'flow': str(flow.uuid)})
+        force_authenticate(request, user=self.user)
+        response = view(request)
+        self.assertEqual(response.status_code, HTTP_200_OK)
+        self.assertTrue(OnboardingSession.objects.filter(flow=flow, user=self.user).exists())
+
+    def test_get_or_create_session_reuses_active(self):
+        flow = OnboardingFlow.objects.create(role=self.role, agent=self.agent, title='Flow', description='')
+        existing = OnboardingSession.objects.create(flow=flow, user=self.user, current_page_order=1)
+        view = OnboardingSessionViewSet.as_view({'post': 'get_or_create'})
+        request = self.factory.post('/', {'flow': str(flow.uuid)})
+        force_authenticate(request, user=self.user)
+        response = view(request)
+        self.assertEqual(response.status_code, HTTP_200_OK)
+        self.assertEqual(response.data.get('uuid'), str(existing.uuid))
+        self.assertEqual(response.data.get('current_page_order'), 1)
+
+    def test_get_or_create_session_creates_after_completion(self):
+        flow = OnboardingFlow.objects.create(role=self.role, agent=self.agent, title='Flow', description='')
+        completed = OnboardingSession.objects.create(flow=flow, user=self.user, status='completed')
+        view = OnboardingSessionViewSet.as_view({'post': 'get_or_create'})
+        request = self.factory.post('/', {'flow': str(flow.uuid)})
+        force_authenticate(request, user=self.user)
+        response = view(request)
+        self.assertEqual(response.status_code, HTTP_200_OK)
+        self.assertNotEqual(response.data.get('uuid'), str(completed.uuid))
+
+

apps/onboarding/tests/test_models.py

@@ -0,0 +1,41 @@
+from django.test import TestCase
+from django.contrib.auth import get_user_model
+from apps.orgs.models import Organization, Role
+from apps.mlstore.models import AgentModel, Agent
+from apps.onboarding.models import OnboardingFlow, OnboardingPage, OnboardingField, OnboardingSession
+
+User = get_user_model()
+
+class OnboardingModelTests(TestCase):
+    def setUp(self):
+        self.user = User.objects.create_user(email_address='user@example.com', password='pass')
+        self.manager = User.objects.create_user(email_address='manager@example.com', password='pass', is_manager=True)
+        self.org = Organization.objects.create(name='Org', owner=self.manager)
+        self.role = Role.objects.create(name='Engineer', organization=self.org)
+        self.model = AgentModel.objects.create(name='test-model', version='v1', path='model.gguf')
+        self.agent = Agent.objects.create(model=self.model, organization=self.org)
+
+    def test_flow_str(self):
+        flow = OnboardingFlow.objects.create(role=self.role, agent=self.agent, title='Welcome', description='Intro')
+        self.assertIn('Welcome', str(flow))
+        self.assertIn(self.role.name, str(flow))
+
+    def test_page_and_field_str(self):
+        flow = OnboardingFlow.objects.create(role=self.role, agent=self.agent, title='Flow', description='')
+        page = OnboardingPage.objects.create(flow=flow, order=0, title='Page 1', body='Body')
+        field = OnboardingField.objects.create(page=page, order=0, key='q1', label='Question 1')
+        self.assertIn(flow.title, str(page))
+        self.assertIn(field.label, str(field))
+
+    def test_field_unique_key_per_page(self):
+        flow = OnboardingFlow.objects.create(role=self.role, agent=self.agent, title='Flow', description='')
+        page = OnboardingPage.objects.create(flow=flow, order=0, title='Page 1', body='Body')
+        OnboardingField.objects.create(page=page, order=0, key='dup', label='Dup 1')
+        with self.assertRaises(Exception):
+            OnboardingField.objects.create(page=page, order=1, key='dup', label='Dup 2')
+
+    def test_session_str(self):
+        flow = OnboardingFlow.objects.create(role=self.role, agent=self.agent, title='Flow', description='')
+        session = OnboardingSession.objects.create(flow=flow, user=self.user)
+        self.assertIn(self.user.email_address, str(session))
+        self.assertIn(flow.title, str(session))

apps/onboarding/viewsets.py

@@ -0,0 +1,451 @@
+import json
+import logging
+import re
+import html
+from typing import Any
+from django.db import transaction
+from django.utils import timezone
+from rest_framework import status
+from rest_framework.exceptions import PermissionDenied
+from rest_framework.decorators import action
+from rest_framework.response import Response
+from rest_framework.viewsets import ModelViewSet
+from asgiref.sync import async_to_sync
+from channels.layers import get_channel_layer
+
+from apps.mlstore.models import AgentEvent, AgentRun
+from apps.mlstore import services as ml_services
+from .models import OnboardingFlow, OnboardingPage, OnboardingField, OnboardingSession
+from .serializers import (
+    OnboardingFlowSerializer,
+    OnboardingFlowDetailSerializer,
+    OnboardingPageSerializer,
+    OnboardingFieldSerializer,
+    OnboardingSessionSerializer,
+    OnboardingSubmissionSerializer,
+    OnboardingFeedbackSerializer,
+)
+
+logger = logging.getLogger(__name__)
+
+
+def _extract_json(text: str) -> dict[str, Any]:
+    if not text:
+        return {}
+
+    try:
+        return json.loads(text)
+    except Exception:
+        pass
+
+    # Prefer fenced json blocks
+    fenced = re.search(r"```json\s*(\{[\s\S]*?\})\s*```", text, re.IGNORECASE)
+    if fenced:
+        try:
+            return json.loads(fenced.group(1))
+        except Exception:
+            return {}
+
+    # Fallback: find first balanced JSON object
+    start = text.find('{')
+    if start == -1:
+        return {}
+
+    depth = 0
+    for idx in range(start, len(text)):
+        char = text[idx]
+        if char == '{':
+            depth += 1
+        elif char == '}':
+            depth -= 1
+            if depth == 0:
+                candidate = text[start:idx + 1]
+                try:
+                    return json.loads(candidate)
+                except Exception:
+                    return {}
+
+    return {}
+
+
+def _strip_html(text: str) -> str:
+    if not text:
+        return ""
+    cleaned = re.sub(r"<[^>]+>", " ", text)
+    cleaned = html.unescape(cleaned)
+    return re.sub(r"\s+", " ", cleaned).strip()
+
+
+def _send_agent_progress_event(agent_run: AgentRun, content: dict):
+    try:
+        AgentEvent.objects.create(
+            execution=agent_run,
+            event_type='progress',
+            content=content,
+        )
+        room_group_name = f"mlstore_agent_{agent_run.agent.uuid}"
+        async_to_sync(get_channel_layer().group_send)(
+            room_group_name,
+            {
+                "type": "mlstore_event",
+                "event_type": "progress",
+                "content": content,
+                "timestamp": timezone.now().isoformat(),
+            },
+        )
+    except Exception as e:
+        logger.warning("Failed to send progress event: %s", e)
+
+
+class OnboardingFlowViewSet(ModelViewSet):
+    queryset = OnboardingFlow.objects.select_related('role', 'agent').all()
+    serializer_class = OnboardingFlowSerializer
+    lookup_field = 'uuid'
+
+    def get_queryset(self):
+        qs = super().get_queryset()
+        role_uuid = self.request.query_params.get('role')
+        status_filter = self.request.query_params.get('status')
+        if role_uuid:
+            qs = qs.filter(role__uuid=role_uuid)
+        if status_filter:
+            qs = qs.filter(status=status_filter)
+        return qs
+
+    def get_serializer_class(self):
+        if self.action in ('retrieve', 'pages'):
+            return OnboardingFlowDetailSerializer
+        return super().get_serializer_class()
+
+    @action(detail=True, methods=['get'])
+    def pages(self, request, pk=None, uuid=None):
+        flow = self.get_object()
+        serializer = OnboardingFlowDetailSerializer(flow, context={'request': request})
+        return Response(serializer.data)
+
+    @action(detail=True, methods=['post'])
+    def generate(self, request, pk=None, uuid=None):
+        flow = self.get_object()
+        if not request.user.is_authenticated or not getattr(request.user, 'is_manager', False):
+            return Response({"error": "permission_denied"}, status=status.HTTP_403_FORBIDDEN)
+        if not flow.agent or not flow.agent.model or not flow.agent.model.path:
+            return Response(
+                {"error": "flow_agent_model_required"},
+                status=status.HTTP_400_BAD_REQUEST,
+            )
+
+        instructions = request.data.get('instructions') or ''
+        rag_context = ""
+        try:
+            rag_context = ml_services.get_context_for_query(
+                query=f"Create onboarding content for role {flow.role.name}",
+                role_uuid=str(flow.role.uuid),
+                top_k=6,
+                similarity_threshold=0.35,
+            )
+        except Exception as e:
+            logger.warning("Onboarding generation RAG lookup failed: %s", e)
+        prompt = (
+            "You are creating onboarding content as JSON. "
+            "Return ONLY valid JSON (no prose, no markdown, no code fences).\n"
+            "Do not include explanations or examples.\n"
+            "Do not include HTML tags. Use plain text only.\n"
+            "Each page body must be 3-6 paragraphs, at least 320 words total, and include 1 short list of 3-5 bullets.\n"
+            "Before writing the body, create a brief outline of the key points to cover and include it in meta.outline.\n"
+            "The outline should be a short list of 3-6 bullets, not chain-of-thought.\n"
+            "Do NOT ask about the learner's personal experience. Ask about what someone in the role may encounter.\n"
+            "Do NOT use any select or multiselect fields. Use only text, textarea, number, boolean, or date.\n"
+            "Use the provided context for accurate, role-specific content.\n"
+            "If context is insufficient, make reasonable assumptions without inventing tools or policies.\n"
+            "JSON shape:\n"
+            "{\n"
+            "  \"title\": string,\n"
+            "  \"description\": string,\n"
+            "  \"pages\": [\n"
+            "    {\n"
+            "      \"title\": string,\n"
+            "      \"body\": string,\n"
+            "      \"meta\": { \"outline\": [string] },\n"
+            "      \"fields\": [\n"
+            "        {\n"
+            "          \"key\": string,\n"
+            "          \"label\": string,\n"
+            "          \"type\": one of [text, textarea, number, boolean, date],\n"
+            "          \"required\": boolean,\n"
+            "          \"help_text\": string,\n"
+            "          \"placeholder\": string,\n"
+            "          \"options\": []\n"
+            "        }\n"
+            "      ]\n"
+            "    }\n"
+            "  ]\n"
+            "}\n"
+            f"Role: {flow.role.name}\n"
+            f"Role description: {flow.role.description}\n"
+            f"Flow title: {flow.title}\n"
+            f"Flow description: {flow.description}\n"
+            f"Extra instructions: {instructions}\n"
+            f"Context:\n{rag_context}\n"
+        )
+
+        try:
+            result = ml_services.infer_with_model(flow.agent.model.path, prompt, {
+                "max_tokens": 1800,
+                "temperature": 0.2,
+            })
+        except Exception as e:
+            logger.error("Onboarding generate inference failed: %s", e, exc_info=True)
+            return Response({"error": "generation_failed"}, status=status.HTTP_500_INTERNAL_SERVER_ERROR)
+
+        response_text = ''
+        if isinstance(result, dict):
+            response_text = result.get('response') or result.get('result') or ''
+        payload = _extract_json(str(response_text))
+        if not payload or 'pages' not in payload:
+            return Response({"error": "invalid_generation_output", "raw": response_text}, status=status.HTTP_400_BAD_REQUEST)
+
+        with transaction.atomic():
+            flow.title = payload.get('title') or flow.title
+            # Keep existing description on regenerate unless explicitly empty
+            if not flow.description:
+                flow.description = payload.get('description') or flow.description
+            if flow.status != 'draft':
+                flow.status = 'draft'
+            flow.save(update_fields=['title', 'description', 'status'])
+
+            OnboardingPage.objects.filter(flow=flow).delete()
+
+            pages = payload.get('pages') or []
+            for page_index, page in enumerate(pages):
+                body_text = _strip_html(page.get('body') or '')
+                page_obj = OnboardingPage.objects.create(
+                    flow=flow,
+                    order=page_index,
+                    title=page.get('title') or f"Page {page_index + 1}",
+                    body=body_text,
+                    meta=page.get('meta') or {},
+                )
+                for field_index, field in enumerate(page.get('fields') or []):
+                    field_type = field.get('type') or 'text'
+                    if field_type not in {"text", "textarea", "number", "boolean", "date"}:
+                        field_type = 'text'
+                    OnboardingField.objects.create(
+                        page=page_obj,
+                        order=field_index,
+                        key=field.get('key') or f"field_{field_index + 1}",
+                        label=field.get('label') or f"Field {field_index + 1}",
+                        field_type=field_type,
+                        required=bool(field.get('required')),
+                        help_text=field.get('help_text') or '',
+                        placeholder=field.get('placeholder') or '',
+                        options=[],
+                        default_value=field.get('default_value') if field.get('default_value') is not None else None,
+                        validation=field.get('validation') or {},
+                    )
+
+        serializer = OnboardingFlowDetailSerializer(flow, context={'request': request})
+        return Response(serializer.data)
+
+    @action(detail=True, methods=['post'])
+    def publish(self, request, pk=None, uuid=None):
+        flow = self.get_object()
+        if not request.user.is_authenticated or not getattr(request.user, 'is_manager', False):
+            return Response({"error": "permission_denied"}, status=status.HTTP_403_FORBIDDEN)
+        if flow.status != 'published':
+            flow.status = 'published'
+            flow.save(update_fields=['status'])
+        serializer = OnboardingFlowDetailSerializer(flow, context={'request': request})
+        return Response(serializer.data)
+
+
+class OnboardingPageViewSet(ModelViewSet):
+    queryset = OnboardingPage.objects.select_related('flow').prefetch_related('fields').all()
+    serializer_class = OnboardingPageSerializer
+    lookup_field = 'uuid'
+
+
+class OnboardingFieldViewSet(ModelViewSet):
+    queryset = OnboardingField.objects.select_related('page').all()
+    serializer_class = OnboardingFieldSerializer
+    lookup_field = 'uuid'
+
+
+class OnboardingSessionViewSet(ModelViewSet):
+    queryset = OnboardingSession.objects.select_related('flow', 'user', 'agent_run', 'flow__agent').all()
+    serializer_class = OnboardingSessionSerializer
+    lookup_field = 'uuid'
+
+    def get_queryset(self):
+        qs = super().get_queryset()
+        user = self.request.user
+        if user.is_authenticated and not getattr(user, 'is_manager', False):
+            qs = qs.filter(user=user)
+        return qs
+
+    def perform_create(self, serializer):
+        if not self.request.user or not self.request.user.is_authenticated:
+            raise PermissionDenied("Authentication required")
+        flow = serializer.validated_data.get('flow')
+        agent_run = None
+        if flow and flow.agent:
+            agent_run = AgentRun.objects.create(
+                agent=flow.agent,
+                user=self.request.user,
+                input_data={
+                    "type": "onboarding_session",
+                    "flow_uuid": str(flow.uuid),
+                    "role_uuid": str(flow.role.uuid),
+                },
+            )
+        serializer.save(user=self.request.user, agent_run=agent_run)
+
+    @action(detail=False, methods=['post'])
+    def get_or_create(self, request):
+        if not request.user or not request.user.is_authenticated:
+            raise PermissionDenied("Authentication required")
+
+        flow_uuid = request.data.get('flow')
+        if not flow_uuid:
+            return Response({"error": "flow_required"}, status=status.HTTP_400_BAD_REQUEST)
+
+        try:
+            flow = OnboardingFlow.objects.get(uuid=flow_uuid)
+        except OnboardingFlow.DoesNotExist:
+            return Response({"error": "flow_not_found"}, status=status.HTTP_404_NOT_FOUND)
+
+        session = (
+            OnboardingSession.objects
+            .filter(flow=flow, user=request.user)
+            .exclude(status='completed')
+            .order_by('-updated_at')
+            .first()
+        )
+
+        if not session:
+            agent_run = None
+            if flow.agent:
+                agent_run = AgentRun.objects.create(
+                    agent=flow.agent,
+                    user=request.user,
+                    input_data={
+                        "type": "onboarding_session",
+                        "flow_uuid": str(flow.uuid),
+                        "role_uuid": str(flow.role.uuid),
+                    },
+                )
+            session = OnboardingSession.objects.create(
+                flow=flow,
+                user=request.user,
+                agent_run=agent_run,
+            )
+
+        return Response(OnboardingSessionSerializer(session, context={'request': request}).data)
+
+    @action(detail=True, methods=['post'])
+    def submit(self, request, pk=None, uuid=None):
+        session = self.get_object()
+        serializer = OnboardingSubmissionSerializer(data=request.data)
+        serializer.is_valid(raise_exception=True)
+        page_uuid = serializer.validated_data['page_uuid']
+        responses = serializer.validated_data['responses']
+        mark_complete = serializer.validated_data.get('mark_complete')
+
+        try:
+            page = OnboardingPage.objects.get(flow=session.flow, uuid=page_uuid)
+        except OnboardingPage.DoesNotExist:
+            return Response({"error": "page_not_found"}, status=status.HTTP_404_NOT_FOUND)
+
+        responses_payload = dict(session.responses or {})
+        responses_payload[str(page.uuid)] = responses
+        session.responses = responses_payload
+        session.current_page_order = page.order
+
+        if mark_complete or page.order >= session.flow.pages.count() - 1:
+            session.status = 'completed'
+            session.completed_at = timezone.now()
+        session.save(update_fields=['responses', 'current_page_order', 'status', 'completed_at'])
+
+        if session.agent_run:
+            progress_payload = {
+                "flow_uuid": str(session.flow.uuid),
+                "session_uuid": str(session.uuid),
+                "page_uuid": str(page.uuid),
+                "page_order": page.order,
+                "status": session.status,
+                "responses": responses,
+            }
+            _send_agent_progress_event(session.agent_run, progress_payload)
+            session.agent_run.output_data = {
+                **(session.agent_run.output_data or {}),
+                "onboarding": session.responses,
+            }
+            session.agent_run.save(update_fields=['output_data'])
+
+        return Response(OnboardingSessionSerializer(session, context={'request': request}).data)
+
+    @action(detail=True, methods=['post'])
+    def feedback(self, request, pk=None, uuid=None):
+        session = self.get_object()
+        serializer = OnboardingFeedbackSerializer(data=request.data)
+        serializer.is_valid(raise_exception=True)
+        page_uuid = serializer.validated_data['page_uuid']
+        responses = serializer.validated_data['responses']
+        question = serializer.validated_data.get('question') or ''
+
+        try:
+            page = OnboardingPage.objects.get(flow=session.flow, uuid=page_uuid)
+        except OnboardingPage.DoesNotExist:
+            return Response({"error": "page_not_found"}, status=status.HTTP_404_NOT_FOUND)
+
+        if not session.flow.agent or not session.flow.agent.model or not session.flow.agent.model.path:
+            return Response({"error": "flow_agent_model_required"}, status=status.HTTP_400_BAD_REQUEST)
+
+        prompt = (
+            "You are an onboarding assessor. Provide concise feedback addressed directly to the learner using second-person \"You\" statements.\n"
+            "Return ONLY valid JSON (no prose, no markdown, no code fences).\n"
+            "JSON shape:\n"
+            "{\n"
+            "  \"summary\": string\n"
+            "}\n\n"
+            f"Page title: {page.title}\n"
+            f"Page body: {page.body}\n"
+            f"Responses: {json.dumps(responses)}\n"
+        )
+        if question:
+            prompt += f"Learner question: {question}\n"
+
+        try:
+            result = ml_services.infer_with_model(session.flow.agent.model.path, prompt, {
+                "max_tokens": 900,
+                "temperature": 0.2,
+            })
+        except Exception as e:
+            logger.error("Onboarding feedback inference failed: %s", e, exc_info=True)
+            return Response({"error": "feedback_failed"}, status=status.HTTP_500_INTERNAL_SERVER_ERROR)
+
+        feedback_text = ''
+        if isinstance(result, dict):
+            feedback_text = result.get('response') or result.get('result') or ''
+        feedback_text = str(feedback_text).strip()
+
+        feedback_payload = _extract_json(feedback_text)
+        if not feedback_payload:
+            feedback_payload = {
+                "summary": feedback_text or "Feedback generated.",
+            }
+
+        responses_payload = dict(session.responses or {})
+        feedback_store = dict(responses_payload.get("__feedback__") or {})
+        feedback_store[str(page.uuid)] = {
+            "feedback": feedback_payload,
+            "question": question,
+            "updated_at": timezone.now().isoformat(),
+        }
+        responses_payload["__feedback__"] = feedback_store
+        session.responses = responses_payload
+        session.save(update_fields=['responses'])
+
+        return Response({
+            "feedback": feedback_payload,
+            "session": OnboardingSessionSerializer(session, context={'request': request}).data,
+        })

apps/orgs/admin.py

@@ -54,8 +54,8 @@ class RoleMembershipAdmin(ModelAdmin):
 
 @register(TrainingFile)
 class TrainingFileAdmin(ModelAdmin):
-	list_display = ('id', 'uuid', 'file_name', 'organization', 'uploaded_by', 'is_processed', 'created_at')
+	list_display = ('id', 'uuid', 'file_name', 'role', 'uploaded_by', 'status', 'is_processed', 'created_at')
-	search_fields = ('file_name', 'organization__name', 'uploaded_by__email_address')
+	search_fields = ('file_name', 'role__name', 'uploaded_by__email_address')
-	list_filter = ('is_processed', 'created_at')
+	list_filter = ('status', 'is_processed', 'created_at')
-	raw_id_fields = ('organization', 'uploaded_by')
+	raw_id_fields = ('role', 'uploaded_by')
 	readonly_fields = ('uuid', 'created_at', 'updated_at')

apps/orgs/migrations/0001_initial.py

@@ -1,11 +1,8 @@
-# Generated by Django 5.2.10 on 2026-01-25 11:02
-
 import django.db.models.deletion
 import uuid
 from django.conf import settings
 from django.db import migrations, models
 
-
 class Migration(migrations.Migration):
 
     initial = True
@@ -118,8 +115,9 @@ class Migration(migrations.Migration):
                 ('file_size', models.IntegerField()),
                 ('file_type', models.CharField(max_length=50)),
                 ('description', models.TextField(blank=True, default='')),
+                ('status', models.CharField(choices=[('ingesting', 'Ingesting'), ('chunked', 'Chunked'), ('embedded', 'Embedded'), ('failed', 'Failed')], default='ingesting', max_length=20)),
                 ('is_processed', models.BooleanField(default=False)),
-                ('organization', models.ForeignKey(on_delete=django.db.models.deletion.CASCADE, related_name='training_files', to='orgs.organization')),
+                ('role', models.ForeignKey(on_delete=django.db.models.deletion.CASCADE, related_name='training_files', to='orgs.role')),
                 ('uploaded_by', models.ForeignKey(on_delete=django.db.models.deletion.CASCADE, related_name='uploaded_training_files', to=settings.AUTH_USER_MODEL)),
             ],
             options={

apps/orgs/models.py

@@ -3,7 +3,8 @@ from uuid import uuid4
 from django.utils import timezone
 from django.utils.translation import gettext_lazy as _
 from django.db.models import BigAutoField, BooleanField, CASCADE, CharField, DateTimeField, ForeignKey, ManyToManyField, Model, TextField, UUIDField, IntegerField, FileField
-from django.db.models.signals import post_delete
+from django.db.models.signals import post_delete, post_save
+from django.db import transaction
 from django.dispatch import receiver
 from apps.users.mixins import TimeStampMixin
 from apps.users.models import User
@@ -104,9 +105,16 @@ class TrainingFile(TimeStampMixin, Model):
     
     ALLOWED_EXTENSIONS = ('txt', 'pdf', 'md', 'csv', 'json', 'docx', 'doc')
     
+    STATUS_CHOICES = [
+        ('ingesting', 'Ingesting'),
+        ('chunked', 'Chunked'),
+        ('embedded', 'Embedded'),
+        ('failed', 'Failed'),
+    ]
+
     id           = BigAutoField(primary_key = True)
     uuid         = UUIDField(default = uuid4, unique = True, editable = False)
-    organization = ForeignKey(Organization, on_delete = CASCADE, related_name = "training_files")
+    role         = ForeignKey(Role, on_delete = CASCADE, related_name = "training_files")
     uploaded_by  = ForeignKey(User, on_delete = CASCADE, related_name = "uploaded_training_files")
     
     file         = FileField(upload_to = 'training_files/%Y/%m/%d/')
@@ -115,6 +123,7 @@ class TrainingFile(TimeStampMixin, Model):
     file_type    = CharField(max_length = 50)
     
     description  = TextField(blank = True, default = '')
+    status       = CharField(max_length = 20, choices = STATUS_CHOICES, default = 'ingesting')
     is_processed = BooleanField(default = False)
     
     class Meta:
@@ -123,7 +132,7 @@ class TrainingFile(TimeStampMixin, Model):
         ordering = ['-created_at']
     
     def __str__(self) -> str:
-        return f"{self.file_name} - {self.organization.name}"
+        return f"{self.file_name} - {self.role.name}"
 
 
 @receiver(post_delete, sender=TrainingFile)
@@ -135,3 +144,15 @@ def delete_training_file_on_delete(sender, instance, **kwargs):
                 os.remove(instance.file.path)
         except Exception:
             pass
+
+
+@receiver(post_save, sender=TrainingFile)
+def enqueue_training_file_ingestion(sender, instance, created, **kwargs):
+    if not created:
+        return
+
+    def _enqueue():
+        from apps.mlstore.tasks import ingest_training_file_task
+        ingest_training_file_task.delay(str(instance.uuid))
+
+    transaction.on_commit(_enqueue)

apps/orgs/serializers.py

@@ -1,4 +1,5 @@
-from rest_framework.serializers import ModelSerializer, SerializerMethodField, IntegerField
+from rest_framework.serializers import ModelSerializer, SerializerMethodField, IntegerField, UUIDField
+from rest_framework.exceptions import ValidationError
 from apps.orgs.models import Organization, OrganizationMembership, OrganizationInvitation, Role, RoleMembership, TrainingFile
 from apps.users.serializers import UserSerializer
 
@@ -75,11 +76,13 @@ class RoleSerializer(ModelSerializer):
 class TrainingFileSerializer(ModelSerializer):
     uploaded_by = UserSerializer(read_only = True)
     file_url = SerializerMethodField()
+    role = RoleSerializer(read_only = True)
+    role_uuid = UUIDField(write_only = True, required = True)
 
     class Meta:
         model = TrainingFile
-        fields = ['id', 'uuid', 'organization', 'uploaded_by', 'file', 'file_name', 'file_size', 'file_type', 'description', 'is_processed', 'file_url', 'created_at', 'updated_at']
+        fields = ['id', 'uuid', 'role', 'role_uuid', 'uploaded_by', 'file', 'file_name', 'file_size', 'file_type', 'description', 'status', 'is_processed', 'file_url', 'created_at', 'updated_at']
-        read_only_fields = ['uuid', 'uploaded_by', 'file_size', 'file_type', 'is_processed', 'created_at', 'updated_at', 'organization']
+        read_only_fields = ['uuid', 'uploaded_by', 'file_size', 'file_type', 'status', 'is_processed', 'created_at', 'updated_at', 'role']
 
     def get_file_url(self, obj):
         request = self.context.get('request')
@@ -88,7 +91,6 @@ class TrainingFileSerializer(ModelSerializer):
         return None
 
     def validate_file(self, value):
-        """Validate that uploaded file is a text-based file."""
         if not value:
             raise ValueError('File is required')
         
@@ -108,10 +110,18 @@ class TrainingFileSerializer(ModelSerializer):
         return value
     
     def create(self, validated_data):
+        role_uuid = validated_data.pop('role_uuid', None)
         file_obj = validated_data.get('file')
         if file_obj:
             validated_data['file_size'] = file_obj.size
             import os
             file_extension = os.path.splitext(file_obj.name)[1][1:].lower()
             validated_data['file_type'] = file_extension
+        if not role_uuid:
+            raise ValidationError({'role_uuid': 'Role is required'})
+        try:
+            role = Role.objects.get(uuid = role_uuid)
+        except Role.DoesNotExist:
+            raise ValidationError({'role_uuid': 'Role not found'})
+        validated_data['role'] = role
         return super().create(validated_data)

apps/orgs/viewsets.py

@@ -166,6 +166,12 @@ class OrganizationViewSet(ModelViewSet):
         serializer = RoleSerializer(role)
         return Response(serializer.data)
 
+    @action(detail=False, methods=['get'], url_path='role/mine')
+    def my_roles(self, request):
+        roles = Role.objects.filter(memberships__user=request.user).distinct()
+        serializer = RoleSerializer(roles, many=True)
+        return Response(serializer.data)
+    
     @action(detail=True, methods=['post'], url_path='role/(?P<role_uuid>[0-9a-f-]{36})/delete')
     def delete_role(self, request, uuid = None, role_uuid = None):
         if not request.user.is_manager:
@@ -196,8 +202,11 @@ class OrganizationViewSet(ModelViewSet):
         organization = self.get_object()
         
         if request.method == 'GET':
-            training_files = TrainingFile.objects.filter(organization=organization)
+            role_uuid = request.query_params.get('role_uuid')
-            serializer = TrainingFileSerializer(training_files, many=True)
+            training_files = TrainingFile.objects.filter(role__organization=organization)
+            if role_uuid:
+                training_files = training_files.filter(role__uuid=role_uuid, role__organization=organization)
+            serializer = TrainingFileSerializer(training_files, many=True, context={'request': request})
             return Response(serializer.data)
         
         if not (organization.owner == request.user or 
@@ -207,9 +216,17 @@ class OrganizationViewSet(ModelViewSet):
                 status=HTTP_403_FORBIDDEN
             )
         
-        serializer = TrainingFileSerializer(data=request.data)
+        role_uuid = request.data.get('role_uuid')
+        if not role_uuid:
+            return Response({'error': 'role_uuid is required'}, status=HTTP_400_BAD_REQUEST)
+        try:
+            Role.objects.get(uuid=role_uuid, organization=organization)
+        except Role.DoesNotExist:
+            return Response({'error': 'Role not found in this organization'}, status=HTTP_404_NOT_FOUND)
+
+        serializer = TrainingFileSerializer(data=request.data, context={'request': request})
         if serializer.is_valid():
-            serializer.save(uploaded_by=request.user, organization=organization)
+            serializer.save(uploaded_by=request.user)
             return Response(serializer.data, status=201)
         return Response(serializer.errors, status=HTTP_400_BAD_REQUEST)
     
@@ -218,17 +235,17 @@ class OrganizationViewSet(ModelViewSet):
         organization = self.get_object()
         
         try:
-            training_file = TrainingFile.objects.get(uuid=file_uuid, organization=organization)
+            training_file = TrainingFile.objects.get(uuid=file_uuid, role__organization=organization)
         except TrainingFile.DoesNotExist:
             return Response({'error': 'Training file not found'}, status=HTTP_404_NOT_FOUND)
         
         if request.method == 'GET':
-            serializer = TrainingFileSerializer(training_file)
+            serializer = TrainingFileSerializer(training_file, context={'request': request})
             return Response(serializer.data)
         
         if not (training_file.uploaded_by == request.user or 
-                training_file.organization.owner == request.user or
+            training_file.role.organization.owner == request.user or
-                request.user.is_manager):
+            request.user.is_manager):
             return Response(
                 {'error': 'You do not have permission to delete this file'},
                 status=HTTP_403_FORBIDDEN

config/api.py

@@ -3,11 +3,16 @@ from rest_framework.routers import DefaultRouter
 from apps.orgs.viewsets import OrganizationViewSet
 from apps.users.viewsets import UserViewSet
 from apps.mlstore.viewsets import AgentViewSet, AgentRunViewSet
+from apps.onboarding.viewsets import OnboardingFlowViewSet, OnboardingPageViewSet, OnboardingFieldViewSet, OnboardingSessionViewSet
 
 router = DefaultRouter()
 router.register(r'user', UserViewSet, basename='user')
 router.register(r'organization', OrganizationViewSet, basename='organization')
 router.register(r'agent', AgentViewSet, basename='agent')
 router.register(r'agent-run', AgentRunViewSet, basename='agent-run')
+router.register(r'onboarding/flow', OnboardingFlowViewSet, basename='onboarding-flow')
+router.register(r'onboarding/page', OnboardingPageViewSet, basename='onboarding-page')
+router.register(r'onboarding/field', OnboardingFieldViewSet, basename='onboarding-field')
+router.register(r'onboarding/session', OnboardingSessionViewSet, basename='onboarding-session')
 
 urlpatterns = router.urls

config/settings.py

@@ -71,6 +71,7 @@ LOCAL_APPS = [
     'apps.users',
     'apps.orgs',
     'apps.mlstore',
+    'apps.onboarding',
 ]
 INSTALLED_APPS = OVERRIDE_APPS + DJANGO_APPS + THIRD_PARTY_APPS + LOCAL_APPS
 

mcp_agent/mcp_server.py

@@ -11,34 +11,36 @@ from aiohttp import web
 from mcp.server import Server
 from mcp.types import Tool, TextContent
 
-logging.basicConfig(
-    level=logging.INFO,
-    format='%(asctime)s - %(name)s - %(levelname)s - %(message)s',
-    handlers=[
-        logging.StreamHandler(sys.stderr),
-        logging.StreamHandler(sys.stdout),
-    ]
-)
 logger = logging.getLogger(__name__)
 
 project_root = os.path.dirname(os.path.dirname(os.path.abspath(__file__)))
 model_cache_dir = os.path.join(project_root, "model", "base-model")
-os.makedirs(model_cache_dir, exist_ok=True)
+
-os.environ["HF_HOME"] = model_cache_dir
+def _init_runtime():
-logger.info(f"Project root: {project_root}")
+    if not logging.getLogger().handlers:
-logger.info(f"HuggingFace model cache directory set to: {model_cache_dir}")
+        logging.basicConfig(
+            level=logging.INFO,
+            format='%(asctime)s - %(name)s - %(levelname)s - %(message)s',
+            handlers=[
+                logging.StreamHandler(sys.stderr),
+                logging.StreamHandler(sys.stdout),
+            ]
+        )
+    os.makedirs(model_cache_dir, exist_ok=True)
+    os.environ["HF_HOME"] = model_cache_dir
+    logger.info(f"Project root: {project_root}")
+    logger.info(f"HuggingFace model cache directory set to: {model_cache_dir}")
 
 app = Server("mlstore-mcp-server")
-logger.info("MCP Server initialized: mlstore-mcp-server")
 
 LOADED_MODELS: Dict[str, Dict[str, Any]] = {}
 
 PAIR_EXTRACTOR: Dict[str, Any] = {}
+EMBEDDING_MODEL: Dict[str, Any] = {}
 
 BASE_MODEL_CACHE_DIR = model_cache_dir
 
 
-
 @app.list_tools()
 async def list_tools():
     logger.info("Listing available tools")
@@ -93,6 +95,18 @@ async def list_tools():
                 "required": ["model_path", "prompt"]
             },
         ),
+        Tool(
+            name="embed",
+            description="Generate embeddings for a list of texts",
+            inputSchema={
+                "type": "object",
+                "properties": {
+                    "texts": {"type": "array", "items": {"type": "string"}},
+                    "model": {"type": "string"}
+                },
+                "required": ["texts"]
+            },
+        ),
     ]
     logger.info(f"Available tools: {[t.name for t in tools]}")
     return tools
@@ -110,6 +124,112 @@ def _safe_dir_name(name: str) -> str:
     return "".join(c for c in name if c.isalnum() or c in ("-", "_", ".")).strip(".")
 
 
+def _map_gguf_repo(model_name: str) -> tuple[str | None, str | None]:
+    gguf_repos = {
+        "Llama-3.1-8B-Instruct.gguf": ("bartowski/Meta-Llama-3.1-8B-Instruct-GGUF", "Meta-Llama-3.1-8B-Instruct-Q4_K_M.gguf"),
+        "Meta-Llama-3.1-8B-Instruct.gguf": ("bartowski/Meta-Llama-3.1-8B-Instruct-GGUF", "Meta-Llama-3.1-8B-Instruct-Q4_K_M.gguf"),
+        "Meta-Llama-3.1-8B-Instruct.Q4_0.gguf": ("bartowski/Meta-Llama-3.1-8B-Instruct-GGUF", "Meta-Llama-3.1-8B-Instruct-Q4_0.gguf"),
+        "Llama-3.1-8B-Instruct.Q4_0.gguf": ("bartowski/Meta-Llama-3.1-8B-Instruct-GGUF", "Meta-Llama-3.1-8B-Instruct-Q4_0.gguf"),
+        "Meta-Llama-3-8B-Instruct.Q4_0.gguf": ("bartowski/Meta-Llama-3-8B-Instruct-GGUF", "Meta-Llama-3-8B-Instruct-Q4_0.gguf"),
+        "Llama-3-8B-Instruct.Q4_0.gguf": ("bartowski/Meta-Llama-3-8B-Instruct-GGUF", "Meta-Llama-3-8B-Instruct-Q4_0.gguf"),
+        "mistral-7b-instruct-v0.3.Q4_0.gguf": ("bartowski/Mistral-7B-Instruct-v0.3-GGUF", "Mistral-7B-Instruct-v0.3-Q4_0.gguf"),
+        "Mistral-7B-Instruct-v0.3.Q4_0.gguf": ("bartowski/Mistral-7B-Instruct-v0.3-GGUF", "Mistral-7B-Instruct-v0.3-Q4_0.gguf"),
+    }
+
+    if model_name in gguf_repos:
+        return gguf_repos[model_name]
+
+    base_name = model_name.lower()
+    if "llama" in base_name and "3.1" in base_name and "8b" in base_name:
+        repo_id = "bartowski/Meta-Llama-3.1-8B-Instruct-GGUF"
+        if ".q4_0" in base_name:
+            return repo_id, "Meta-Llama-3.1-8B-Instruct-Q4_0.gguf"
+        if ".q4_k_m" in base_name:
+            return repo_id, "Meta-Llama-3.1-8B-Instruct-Q4_K_M.gguf"
+        return repo_id, "Meta-Llama-3.1-8B-Instruct-Q4_K_M.gguf"
+
+    if "llama" in base_name and "3" in base_name and "8b" in base_name:
+        repo_id = "bartowski/Meta-Llama-3-8B-Instruct-GGUF"
+        if ".q4_0" in base_name:
+            return repo_id, "Meta-Llama-3-8B-Instruct-Q4_0.gguf"
+        return repo_id, "Meta-Llama-3-8B-Instruct-Q4_K_M.gguf"
+
+    if "mistral" in base_name and "7b" in base_name:
+        repo_id = "bartowski/Mistral-7B-Instruct-v0.3-GGUF"
+        if ".q4_0" in base_name:
+            return repo_id, "Mistral-7B-Instruct-v0.3-Q4_0.gguf"
+        return repo_id, "Mistral-7B-Instruct-v0.3-Q4_K_M.gguf"
+
+    return None, None
+
+
+def _find_existing_gguf(model_name: str) -> str | None:
+    repo_id, filename = _map_gguf_repo(model_name)
+    if not filename:
+        return None
+
+    candidate_paths = [
+        os.path.join(_model_root(), filename),
+        os.path.join(model_cache_dir, filename),
+        os.path.join(os.getcwd(), "model", filename),
+        os.path.join(os.getcwd(), filename),
+    ]
+
+    for path in candidate_paths:
+        if os.path.exists(path):
+            return path
+
+    return None
+
+
+def _download_gguf_from_hf(model_name: str) -> str:
+    """
+    Download a GGUF model from Hugging Face Hub.
+    Returns the path to the downloaded model file.
+    """
+    logger.info(f"Attempting to download GGUF model from Hugging Face: {model_name}")
+
+    existing_path = _find_existing_gguf(model_name)
+    if existing_path:
+        logger.info(f"Found existing GGUF locally: {existing_path}")
+        return existing_path
+    
+    try:
+        from huggingface_hub import hf_hub_download, list_repo_files
+        
+        model_dir = _model_root()
+        os.makedirs(model_dir, exist_ok=True)
+        
+        repo_id, filename = _map_gguf_repo(model_name)
+        if repo_id and filename:
+            logger.info(f"Found known model mapping: {repo_id}/{filename}")
+        
+        if not repo_id or not filename:
+            logger.error(f"Could not determine Hugging Face repo for model: {model_name}")
+            raise ValueError(f"Unknown model: {model_name}. Please specify a known GGUF model.")
+        
+        logger.info(f"Downloading {filename} from {repo_id}...")
+        logger.info(f"This may take several minutes depending on model size and connection speed.")
+        
+        downloaded_path = hf_hub_download(
+            repo_id=repo_id,
+            filename=filename,
+            cache_dir=model_cache_dir,
+            local_dir=model_dir,
+            local_dir_use_symlinks=False,
+        )
+        
+        logger.info(f"Model downloaded successfully to: {downloaded_path}")
+        return downloaded_path
+        
+    except ImportError as e:
+        logger.error(f"huggingface_hub not available: {str(e)}")
+        raise ImportError("huggingface_hub is required to download models. Install with: pip install huggingface_hub")
+    except Exception as e:
+        logger.error(f"Failed to download model from Hugging Face: {str(e)}", exc_info=True)
+        raise
+
+
 def _resolve_model_path(model_path: str) -> str:
     if not model_path:
         return model_path
@@ -298,14 +418,16 @@ def _prompt_based_pair_extraction(training_data: List[Any], base_model: str) ->
     system_prompt = (
         "You are a data extractor. Given a list of items, return a JSON array of training pairs. "
         "Each pair must have 'instruction' and 'response'. Keep answers concise. "
-        "If content is incomplete, still produce best-effort pairs."
+        "If content is incomplete, still produce best-effort pairs. "
+        "End your answer with a complete sentence. Do not start lists or new sections."
     )
 
     user_prompt = (
         "Examples of desired output:\n"
         f"{json.dumps(example_pairs, ensure_ascii=False, indent=2)}\n\n"
         "Now extract training pairs from the following items. Return ONLY a JSON array, no extra text.\n"
-        f"Items:\n{data_block}"
+        f"Items:\n{data_block}\n\n"
+        "Answer:"
     )
 
     prompt = f"<|im_start|>system\n{system_prompt}<|im_end|>\n<|im_start|>user\n{user_prompt}<|im_end|>\n<|im_start|>assistant\n"
@@ -385,7 +507,7 @@ def _extract_training_pairs(training_data: List[Any]) -> List[Tuple[str, str]]:
     if not training_data:
         return []
 
-    base_model = "meta-llama/Meta-Llama-3-8B-Instruct"
+    base_model = "meta-llama/Llama-3.1-8B-Instruct"
 
     pairs = _prompt_based_pair_extraction(training_data, base_model)
     if not pairs:
@@ -402,6 +524,19 @@ def _extract_training_pairs(training_data: List[Any]) -> List[Tuple[str, str]]:
     return pairs
 
 
+def _ensure_embedding_model(model_name: str):
+    if EMBEDDING_MODEL.get("model") is not None and EMBEDDING_MODEL.get("name") == model_name:
+        return EMBEDDING_MODEL["model"]
+
+    from sentence_transformers import SentenceTransformer
+
+    logger.info(f"Loading embedding model: {model_name}")
+    model = SentenceTransformer(model_name, cache_folder=model_cache_dir)
+    EMBEDDING_MODEL["model"] = model
+    EMBEDDING_MODEL["name"] = model_name
+    return model
+
+
 async def _fine_tune_model_impl(
     training_files: List[str],
     hyperparams: Dict[str, Any],
@@ -409,7 +544,7 @@ async def _fine_tune_model_impl(
     version: str,
     output_dir: str
 ) -> Dict[str, Any]:
-    base_model = "meta-llama/Meta-Llama-3-8B-Instruct"
+    base_model = "meta-llama/Llama-3.1-8B-Instruct"
     logger.info(f"Starting fine-tune process with base model: {base_model}")
     
     try:
@@ -793,14 +928,47 @@ async def _run_tool_http(name: str, arguments: dict) -> Dict[str, Any]:
             logger.error("model_path_required error: no model path provided")
             return {"status": "failed", "error": "model_path_required", "timestamp": _now()}
 
+        original_model_path = model_path
         model_path = _resolve_model_path(model_path)
         logger.debug(f"Resolved model path: {model_path}")
 
         if not os.path.exists(model_path):
-            logger.error(f"Model not found at: {model_path}")
+            logger.warning(f"Model not found at: {model_path}")
-            return {"status": "failed", "error": "model_not_found", "model_path": model_path, "timestamp": _now()}
+            local_mapped = _find_existing_gguf(original_model_path)
+            if local_mapped:
+                logger.info(f"Using existing mapped GGUF: {local_mapped}")
+                model_path = local_mapped
+            else:
+                logger.info(f"Attempting to download model from Hugging Face...")
+                try:
+                    model_path = _download_gguf_from_hf(original_model_path)
+                    logger.info(f"Model downloaded successfully: {model_path}")
+                except Exception as download_error:
+                    logger.error(f"Failed to download model: {str(download_error)}")
+                    return {
+                        "status": "failed",
+                        "error": "model_not_found_and_download_failed",
+                        "model_path": original_model_path,
+                        "download_error": str(download_error),
+                        "timestamp": _now()
+                    }
 
         try:
+            for loaded_path in list(LOADED_MODELS.keys()):
+                if loaded_path != model_path:
+                    logger.info(f"Unloading cached model: {loaded_path}")
+                    LOADED_MODELS.pop(loaded_path, None)
+
+            if model_path in LOADED_MODELS and "model" in LOADED_MODELS[model_path]:
+                logger.info(f"Model already loaded: {model_path}")
+                return {
+                    "status": "completed",
+                    "model_path": model_path,
+                    "loaded": True,
+                    "cached": True,
+                    "timestamp": _now(),
+                }
+
             from gpt4all import GPT4All
 
             model_dir, model_file = _resolve_model_file(model_path)
@@ -866,16 +1034,38 @@ async def _run_tool_http(name: str, arguments: dict) -> Dict[str, Any]:
             logger.error("model_path_required error: no model path provided")
             return {"status": "failed", "error": "model_path_required", "timestamp": _now()}
 
+        original_model_path = model_path
         model_path = _resolve_model_path(model_path)
         logger.debug(f"Resolved model path: {model_path}")
 
         if not os.path.exists(model_path):
-            logger.error(f"Model not found at: {model_path}")
+            logger.warning(f"Model not found at: {model_path}")
-            return {"status": "failed", "error": "model_not_found", "model_path": model_path, "timestamp": _now()}
+            local_mapped = _find_existing_gguf(original_model_path)
+            if local_mapped:
+                logger.info(f"Using existing mapped GGUF: {local_mapped}")
+                model_path = local_mapped
+            else:
+                logger.info(f"Attempting to download model from Hugging Face...")
+                try:
+                    model_path = _download_gguf_from_hf(original_model_path)
+                    logger.info(f"Model downloaded successfully: {model_path}")
+                except Exception as download_error:
+                    logger.error(f"Failed to download model: {str(download_error)}")
+                    return {
+                        "status": "failed",
+                        "error": "model_not_found_and_download_failed",
+                        "model_path": original_model_path,
+                        "download_error": str(download_error),
+                        "timestamp": _now()
+                    }
 
         try:
             if model_path not in LOADED_MODELS or "model" not in LOADED_MODELS[model_path]:
                 logger.info(f"Model not in memory, loading: {model_path}")
+                for loaded_path in list(LOADED_MODELS.keys()):
+                    if loaded_path != model_path:
+                        logger.info(f"Unloading cached model: {loaded_path}")
+                        LOADED_MODELS.pop(loaded_path, None)
                 from gpt4all import GPT4All
 
                 model_dir, model_file = _resolve_model_file(model_path)
@@ -952,6 +1142,34 @@ async def _run_tool_http(name: str, arguments: dict) -> Dict[str, Any]:
                 "timestamp": _now(),
             }
 
+    if name == "embed":
+        texts = arguments.get("texts") or []
+        model_name = arguments.get("model") or "all-MiniLM-L6-v2"
+
+        if not isinstance(texts, list) or not all(isinstance(t, str) for t in texts):
+            return {"status": "failed", "error": "texts_must_be_list_of_strings", "timestamp": _now()}
+
+        if not texts:
+            return {"status": "completed", "embeddings": [], "timestamp": _now()}
+
+        try:
+            model = _ensure_embedding_model(model_name)
+            embeddings = model.encode(texts).tolist()
+            return {
+                "status": "completed",
+                "embeddings": embeddings,
+                "model": model_name,
+                "timestamp": _now(),
+            }
+        except Exception as e:
+            logger.error(f"Embedding failed: {str(e)}", exc_info=True)
+            return {
+                "status": "failed",
+                "error": str(e),
+                "error_type": type(e).__name__,
+                "timestamp": _now(),
+            }
+
     raise ValueError(f"Unknown tool: {name}")
 
 
@@ -1009,6 +1227,7 @@ async def handle_health(request: web.Request) -> web.Response:
 
 
 async def run_http_server():
+    _init_runtime()
     host = os.getenv("MCP_HTTP_HOST", "0.0.0.0")
     port = int(os.getenv("MCP_HTTP_PORT", "8001"))
     logger.info(f"Starting HTTP server on {host}:{port}")
@@ -1028,6 +1247,7 @@ async def run_http_server():
 
 
 if __name__ == "__main__":
+    _init_runtime()
     logger.info("Starting MCP Server...")
     try:
         asyncio.run(run_http_server())

src/App.vue

@@ -5,9 +5,7 @@ import {
     HomeOutlined,
     InfoCircleOutlined,
     RocketOutlined,
-    TeamOutlined,
     RobotOutlined,
-    BulbOutlined,
     AppstoreOutlined,
     DashboardOutlined,
     LoginOutlined,
@@ -42,10 +40,8 @@ const navItems: NavItem[] = [
         icon: BuildOutlined,
         path: '/organization',
         children: [
-            { key: '/roles', label: 'Roles', icon: TeamOutlined, path: '/roles', manager: true },
             { key: '/onboarding', label: 'Onboarding', icon: RocketOutlined, path: '/onboarding' },
             { key: '/progress', label: 'Progress', icon: DashboardOutlined, path: '/progress' },
-            { key: '/assessments', label: 'Assessments', icon: BulbOutlined, path: '/assessments' },
             { key: '/resources', label: 'Resources', icon: AppstoreOutlined, path: '/resources' },
         ],
     },

src/router/api.ts

@@ -77,6 +77,7 @@ export const API = {
     organizations: () => '/api/organization/',
     organization: (id: string) => `/api/organization/${id}/`,
     organizationRoles: (orgUuid: string) => `/api/organization/${orgUuid}/role/`,
+    organizationRolesMine: () => '/api/organization/role/mine/',
     organizationRole: (orgUuid: string, roleUuid: string) =>
         `/api/organization/${orgUuid}/role/${roleUuid}/`,
     organizationRoleMembers: (orgUuid: string, roleUuid: string) =>
@@ -98,6 +99,18 @@ export const API = {
         `/api/organization/${orgUuid}/training-file/${fileUuid}/`,
     agents: () => '/api/agent/',
     agent: (id: string) => `/api/agent/${id}/`,
+    onboardingFlows: () => '/api/onboarding/flow/',
+    onboardingFlow: (flowUuid: string) => `/api/onboarding/flow/${flowUuid}/`,
+    onboardingFlowPages: (flowUuid: string) => `/api/onboarding/flow/${flowUuid}/pages/`,
+    onboardingFlowGenerate: (flowUuid: string) => `/api/onboarding/flow/${flowUuid}/generate/`,
+    onboardingFlowPublish: (flowUuid: string) => `/api/onboarding/flow/${flowUuid}/publish/`,
+    onboardingSessions: () => '/api/onboarding/session/',
+    onboardingSessionGetOrCreate: () => '/api/onboarding/session/get_or_create/',
+    onboardingSession: (sessionUuid: string) => `/api/onboarding/session/${sessionUuid}/`,
+    onboardingSessionSubmit: (sessionUuid: string) =>
+        `/api/onboarding/session/${sessionUuid}/submit/`,
+    onboardingSessionFeedback: (sessionUuid: string) =>
+        `/api/onboarding/session/${sessionUuid}/feedback/`,
 }
 
 export const apiClient = new ApiClient()

src/router/index.ts

@@ -67,6 +67,18 @@ const router = createRouter({
             component: () => import('../views/AgentDetailView.vue'),
             meta: { requiresAuth: true, requiresManager: true },
         },
+        {
+            path: '/onboarding',
+            name: 'onboarding',
+            component: () => import('../views/OnboardingView.vue'),
+            meta: { requiresAuth: true },
+        },
+        {
+            path: '/onboarding/:roleId',
+            name: 'onboarding-role',
+            component: () => import('../views/OnboardingView.vue'),
+            meta: { requiresAuth: true },
+        },
     ],
 })
 

src/stores/agentStore.ts

@@ -97,15 +97,26 @@ export const useAgentStore = defineStore('agent', () => {
         isConnected.value = false
     }
 
-    const startAgent = (data: { query?: string; prompt?: string; options?: Record<string, unknown> }) => {
+    const startAgent = (data: {
+        query?: string
+        prompt?: string
+        role_uuid?: string
+        max_tokens?: number
+        options?: Record<string, unknown>
+    }) => {
         if (!socket || socket.readyState !== WebSocket.OPEN) return
         const prompt = data.query ?? data.prompt ?? ''
+        const options = {
+            ...(data.options ?? {}),
+            ...(typeof data.max_tokens === 'number' ? { max_tokens: data.max_tokens } : {}),
+        }
         socket.send(
             JSON.stringify({
                 action: 'infer',
                 input_data: {
                     prompt,
-                    options: data.options ?? {},
+                    role_uuid: data.role_uuid,
+                    options,
                 },
             }),
         )
@@ -131,6 +142,17 @@ export const useAgentStore = defineStore('agent', () => {
         )
     }
 
+    const sendOnboardingProgress = (executionId: string, content: Record<string, unknown>) => {
+        if (!socket || socket.readyState !== WebSocket.OPEN) return
+        socket.send(
+            JSON.stringify({
+                action: 'onboarding_progress',
+                execution_id: executionId,
+                content,
+            }),
+        )
+    }
+
     const resetLog = () => {
         eventLog.value = []
     }
@@ -144,6 +166,7 @@ export const useAgentStore = defineStore('agent', () => {
         startAgent,
         startFineTune,
         stopAgent,
+        sendOnboardingProgress,
         resetLog,
         lastExecutionId,
     }

src/types/onboarding.ts

@@ -0,0 +1,44 @@
+export type OnboardingField = {
+    uuid: string
+    key: string
+    label: string
+    field_type: 'text' | 'textarea' | 'select' | 'multiselect' | 'number' | 'boolean' | 'date'
+    required: boolean
+    help_text?: string
+    placeholder?: string
+    options?: string[]
+    default_value?: unknown
+    validation?: Record<string, unknown>
+}
+
+export type OnboardingPage = {
+    uuid: string
+    order: number
+    title: string
+    body?: string
+    meta?: Record<string, unknown>
+    fields: OnboardingField[]
+}
+
+export type OnboardingFlow = {
+    uuid: string
+    role: string
+    agent?: string | null
+    title: string
+    description?: string
+    status: 'draft' | 'published' | 'archived'
+    pages?: OnboardingPage[]
+}
+
+export type OnboardingSession = {
+    uuid: string
+    flow: string
+    agent_run?: string | null
+    status: 'in_progress' | 'completed' | 'abandoned'
+    current_page_order: number
+    responses: Record<string, unknown>
+}
+
+export type OnboardingFeedback = {
+    summary?: string
+}

src/types/organization.ts

@@ -37,7 +37,7 @@ export interface InviteToken {
 export interface TrainingFile {
     id: number
     uuid: string
-    organization: string
+    role: Role
     uploaded_by: User
     file: string
     file_name: string
@@ -45,6 +45,7 @@ export interface TrainingFile {
     file_type: string
     description: string
     is_processed: boolean
+    status: 'ingesting' | 'chunked' | 'embedded' | 'failed'
     file_url: string
     created_at: string
     updated_at: string

src/views/AboutView.vue

@@ -2,15 +2,15 @@
 import { Card, Typography, Divider, List } from 'ant-design-vue'
 
 const steps = [
-    'Register or login (demo credentials only).',
+    'Register or login.',
-    'Complete Onboarding and Training to simulate a role journey.',
+    'Complete onboarding and training to simulate a role journey.',
-    'Managers assign employees to roles and review progress reports.',
+    'Managers review onboarding completion and feedback.',
 ]
 
 const features = [
     {
         title: 'Modular Content',
-        desc: 'Compose learning journeys from small, reusable modules — mix assessments, videos and interactive checks.',
+        desc: 'Compose learning journeys from small, reusable modules. Mix videos and interactive checks.',
         img: 'https://placehold.co/600x400?text=Modular+Content',
     },
     {
@@ -20,7 +20,7 @@ const features = [
     },
     {
         title: 'Reporting & Insights',
-        desc: 'Lightweight reports showing completion, assessment scores and engagement metrics.',
+        desc: 'Lightweight reports showing completion and engagement metrics.',
         img: 'https://placehold.co/600x400?text=Reporting',
     },
 ]
@@ -33,7 +33,7 @@ const features = [
             <Typography.Paragraph type="secondary">
                 Dynavera is a lightweight platform for onboarding, training, and assessing employees
                 with modular content and agent-driven workflows. It is designed for teams that want
-                angible learning experiences quickly without complex LMS setup.
+                tangible learning experiences quickly without complex LMS setup.
             </Typography.Paragraph>
             <Divider />
             <Typography.Title :level="4">Getting started</Typography.Title>
@@ -59,11 +59,11 @@ const features = [
             <Divider />
             <Typography.Title :level="4">More about Dynavera</Typography.Title>
             <Typography.Paragraph>
-                Dynavera is built to be extensible — plug your content sources, connect an LMS, or
+                Dynavera is built to be extensible. Plug your content sources, connect an LMS, or
-                integrate third-party assessment engines. The platform focuses on flexibility and
+                integrate third-party learning tools. The platform focuses on flexibility and ease
-                ease of use, so teams can get started quickly and adapt as their needs evolve.
+                of use, so teams can get started quickly and adapt as their needs evolve. Whether
-                Whether you’re a small startup or a growing enterprise, Dynavera aims to simplify
+                you are a small startup or a growing enterprise, Dynavera aims to simplify the
-                the process of onboarding and training with modern, agent-driven approaches.
+                process of onboarding and training with modern, agent-driven approaches.
             </Typography.Paragraph>
         </Card>
     </div>

src/views/AgentDetailView.vue

@@ -1,9 +1,10 @@
 <script setup lang="ts">
 import { ref, onMounted, onUnmounted, computed } from 'vue'
 import { useRoute } from 'vue-router'
-import { Card, Typography, Button, List, Space, Spin, Input, message, Tag } from 'ant-design-vue'
+import { Card, Typography, Button, List, Space, Spin, Input, message, Tag, Select, InputNumber } from 'ant-design-vue'
 import { useAgentStore } from '../stores/agentStore'
 import { apiClient, isAxiosError, API } from '../router/api'
+import type { Role } from '../types/organization'
 
 const route = useRoute()
 const agentStore = useAgentStore()
@@ -17,6 +18,10 @@ const agent = ref<Record<string, unknown>>({
     status: 'idle',
 })
 
+const roles = ref<Role[]>([])
+const selectedRoleUuid = ref('')
+const maxTokens = ref<number>(256)
+
 const queryInput = ref('')
 const isRunning = computed(() => agentStore.executionStatus === 'running')
 const isConnected = computed(() => agentStore.isConnected ?? false)
@@ -52,6 +57,16 @@ const fetchAgent = async () => {
     try {
         const response = await apiClient.get<Record<string, unknown>>(API.agent(agentId))
         agent.value = response.data
+        const org = agent.value.organization as Record<string, unknown> | null | undefined
+        const orgUuid = org?.uuid as string | undefined
+        console.log('Agent loaded:', agent.value)
+        console.log('Organization:', org)
+        console.log('Organization UUID:', orgUuid)
+        if (orgUuid) {
+            await fetchRoles(orgUuid)
+        } else {
+            console.warn('No organization UUID found for agent')
+        }
     } catch (error) {
         console.error('Failed to fetch agent:', error)
         if (isAxiosError(error)) {
@@ -65,6 +80,32 @@ const fetchAgent = async () => {
     }
 }
 
+const fetchRoles = async (orgUuid: string) => {
+    try {
+        console.log('Fetching roles for organization:', orgUuid)
+        const response = await apiClient.get<Role[]>(API.organizationRoles(orgUuid))
+        console.log('Roles loaded:', response.data)
+        roles.value = response.data
+        if (!selectedRoleUuid.value && roles.value.length > 0) {
+            selectedRoleUuid.value = roles.value[0].uuid
+            console.log('Auto-selected role:', selectedRoleUuid.value)
+        }
+    } catch (error) {
+        console.error('Failed to fetch roles:', error)
+        if (isAxiosError(error)) {
+            console.error('Roles API error:', {
+                status: error.response?.status,
+                data: error.response?.data,
+            })
+        }
+        message.error('Failed to load roles')
+    }
+}
+
+const roleOptions = computed(() =>
+    roles.value.map((role) => ({ label: role.name, value: role.uuid })),
+)
+
 const startAgent = () => {
     if (!agentStore.isConnected) {
         message.error('WebSocket not connected')
@@ -76,8 +117,15 @@ const startAgent = () => {
         return
     }
 
+    if (!selectedRoleUuid.value) {
+        message.error('Please select a role for this query')
+        return
+    }
+
     const data = {
         query: queryInput.value.trim(),
+        role_uuid: selectedRoleUuid.value,
+        max_tokens: maxTokens.value,
     }
 
     agentStore.startAgent(data)
@@ -144,6 +192,28 @@ onUnmounted(() => {
                         />
                     </div>
 
+                    <div>
+                        <Typography.Text>Role:</Typography.Text>
+                        <Select
+                            v-model:value="selectedRoleUuid"
+                            :options="roleOptions"
+                            :disabled="isRunning"
+                            placeholder="Select a role"
+                            style="width: 100%"
+                        />
+                    </div>
+
+                    <div>
+                        <Typography.Text>Max Tokens:</Typography.Text>
+                        <InputNumber
+                            v-model:value="maxTokens"
+                            :min="1"
+                            :max="4096"
+                            :disabled="isRunning"
+                            style="width: 100%"
+                        />
+                    </div>
+
                     <Space>
                         <Button type="primary" :disabled="isRunning || !isConnected" @click="startAgent">
                             Run Agent

src/views/OnboardingView.vue

@@ -0,0 +1,722 @@
+<script setup lang="ts">
+import { ref, computed, onMounted, onUnmounted, watch } from 'vue'
+import type { Dayjs } from 'dayjs'
+import { useRoute, useRouter } from 'vue-router'
+import {
+    Card,
+    Typography,
+    Button,
+    Space,
+    Spin,
+    Select,
+    Form,
+    Input,
+    InputNumber,
+    Switch,
+    DatePicker,
+    Divider,
+    List,
+    message,
+} from 'ant-design-vue'
+import { apiClient, API } from '../router/api'
+import { useUserStore } from '../stores/userStore'
+import { useAgentStore } from '../stores/agentStore'
+import type { Role } from '../types/organization'
+import type { OnboardingFlow, OnboardingPage, OnboardingSession, OnboardingFeedback } from '../types/onboarding'
+
+const route = useRoute()
+const router = useRouter()
+const userStore = useUserStore()
+const agentStore = useAgentStore()
+
+const roleId = computed(() => (route.params.roleId as string | undefined) || undefined)
+const flows = ref<OnboardingFlow[]>([])
+type SingleSelectValue = string | number | undefined
+const selectedFlowUuid = ref<SingleSelectValue>(undefined)
+const flowDetails = ref<OnboardingFlow | null>(null)
+const session = ref<OnboardingSession | null>(null)
+const currentPageIndex = ref(0)
+const loading = ref(false)
+const generating = ref(false)
+const loadError = ref<string | null>(null)
+const generateInstructions = ref('')
+const creatingFlow = ref(false)
+const resettingSession = ref(false)
+const publishing = ref(false)
+const agents = ref<Array<{ uuid: string; name: string }>>([])
+const userRoles = ref<Role[]>([])
+const roleLoading = ref(false)
+const roleLoadError = ref<string | null>(null)
+const createFlowForm = ref({
+    title: '',
+    description: '',
+    agent: undefined as SingleSelectValue,
+})
+
+const isManager = computed(() => userStore.isGeneralManager)
+const hasRoleContext = computed(() => Boolean(roleId.value))
+const pages = computed<OnboardingPage[]>(() => flowDetails.value?.pages ?? [])
+const currentPage = computed<OnboardingPage | null>(() => pages.value[currentPageIndex.value] || null)
+const hasNext = computed(() => currentPageIndex.value < pages.value.length - 1)
+const hasPrev = computed(() => currentPageIndex.value > 0)
+
+const formState = ref<Record<string, unknown>>({})
+const feedbackLoading = ref(false)
+const feedbackByPage = ref<Record<string, OnboardingFeedback>>({})
+
+const getFieldValue = (key: string) => formState.value[key]
+const setFieldValue = (key: string, value: unknown) => {
+    formState.value[key] = value
+}
+const toSingleValue = (value: unknown): SingleSelectValue => {
+    if (Array.isArray(value)) {
+        return value.length ? (value[0] as string | number) : undefined
+    }
+    if (typeof value === 'string' || typeof value === 'number') {
+        return value
+    }
+    return undefined
+}
+const toNumberValue = (value: unknown): number | undefined =>
+    typeof value === 'number' ? value : undefined
+const toDateValue = (value: unknown): Dayjs | string | undefined =>
+    typeof value === 'string' || (value as Dayjs | undefined)?.isValid ? (value as Dayjs | string) : undefined
+
+const fetchFlows = async () => {
+    loading.value = true
+    loadError.value = null
+    try {
+        const params: Record<string, string> = {}
+        if (roleId.value) params.role = roleId.value
+        if (!isManager.value) params.status = 'published'
+        const response = await apiClient.get<OnboardingFlow[]>(API.onboardingFlows(), { params })
+        flows.value = Array.isArray(response.data) ? response.data : []
+        if (flows.value.length === 1) {
+            selectedFlowUuid.value = flows.value[0].uuid
+        }
+    } catch (err: unknown) {
+        console.error('Failed to load onboarding flows', err)
+        loadError.value = 'Failed to load onboarding flows'
+        flows.value = []
+    } finally {
+        loading.value = false
+    }
+}
+
+const fetchAgents = async () => {
+    try {
+        const response = await apiClient.get<Array<{ uuid: string; name: string }>>(API.agents())
+        agents.value = Array.isArray(response.data) ? response.data : []
+    } catch (err: unknown) {
+        console.error('Failed to load agents', err)
+        agents.value = []
+    }
+}
+
+const fetchUserRoles = async () => {
+    roleLoading.value = true
+    roleLoadError.value = null
+    try {
+        const response = await apiClient.get<Role[]>(API.organizationRolesMine())
+        userRoles.value = Array.isArray(response.data) ? response.data : []
+        if (userRoles.value.length === 1) {
+            await router.replace(`/onboarding/${userRoles.value[0].uuid}`)
+        }
+    } catch (err: unknown) {
+        console.error('Failed to load user roles', err)
+        roleLoadError.value = 'Failed to load roles'
+        userRoles.value = []
+    } finally {
+        roleLoading.value = false
+    }
+}
+
+const fetchFlowDetails = async (flowUuid: string) => {
+    loading.value = true
+    loadError.value = null
+    try {
+        const response = await apiClient.get<OnboardingFlow>(API.onboardingFlowPages(flowUuid))
+        flowDetails.value = response.data
+        await ensureSession(flowUuid)
+        syncPageIndexFromSession()
+        hydrateFormState()
+        if (flowDetails.value?.agent) {
+            agentStore.connect(flowDetails.value.agent)
+        }
+    } catch (err: unknown) {
+        console.error('Failed to load onboarding flow', err)
+        loadError.value = 'Failed to load onboarding flow'
+        flowDetails.value = null
+    } finally {
+        loading.value = false
+    }
+}
+
+const ensureSession = async (flowUuid: string) => {
+    if (session.value && session.value.flow === flowUuid) return
+    const response = await apiClient.post<OnboardingSession>(API.onboardingSessionGetOrCreate(), {
+        flow: flowUuid,
+    })
+    session.value = response.data
+}
+
+const syncPageIndexFromSession = () => {
+    const order = session.value?.current_page_order
+    if (typeof order === 'number' && order >= 0) {
+        currentPageIndex.value = Math.min(order, Math.max(pages.value.length - 1, 0))
+    } else {
+        currentPageIndex.value = 0
+    }
+}
+
+const hydrateFormState = () => {
+    if (!currentPage.value) {
+        formState.value = {}
+        return
+    }
+    const existing = session.value?.responses?.[currentPage.value.uuid] || {}
+    const values: Record<string, unknown> = {}
+    currentPage.value.fields.forEach((field) => {
+        const existingValue = (existing as Record<string, unknown>)[field.key]
+        if (existingValue !== undefined) {
+            values[field.key] = existingValue
+        } else if (field.default_value !== undefined && field.default_value !== null) {
+            values[field.key] = field.default_value
+        } else if (field.field_type === 'boolean') {
+            values[field.key] = false
+        } else if (field.field_type === 'multiselect') {
+            values[field.key] = []
+        } else {
+            values[field.key] = ''
+        }
+    })
+    formState.value = values
+
+    const storedFeedback = (session.value?.responses as Record<string, unknown> | undefined)?.[
+        '__feedback__'
+    ] as Record<string, { feedback?: OnboardingFeedback }> | undefined
+    if (storedFeedback && currentPage.value) {
+        const pageFeedback = storedFeedback[currentPage.value.uuid]?.feedback
+        if (pageFeedback) {
+            feedbackByPage.value[currentPage.value.uuid] = pageFeedback
+        }
+    }
+}
+
+const normalizeResponses = (raw: Record<string, unknown>) => {
+    const normalized: Record<string, unknown> = {}
+    Object.entries(raw).forEach(([key, value]) => {
+        if (value && typeof value === 'object') {
+            const maybeDate = value as { toISOString?: () => string; format?: (fmt: string) => string }
+            if (typeof maybeDate.toISOString === 'function') {
+                normalized[key] = maybeDate.toISOString()
+                return
+            }
+            if (typeof maybeDate.format === 'function') {
+                normalized[key] = maybeDate.format('YYYY-MM-DD')
+                return
+            }
+        }
+        normalized[key] = value
+    })
+    return normalized
+}
+
+const onSubmitPage = async () => {
+    if (!currentPage.value || !session.value) return
+    try {
+        const normalizedResponses = normalizeResponses(formState.value)
+        const payload = {
+            page_uuid: currentPage.value.uuid,
+            responses: normalizedResponses,
+            mark_complete: !hasNext.value,
+        }
+        const response = await apiClient.post<OnboardingSession>(
+            API.onboardingSessionSubmit(session.value.uuid),
+            payload,
+        )
+        session.value = response.data
+
+        if (session.value.agent_run) {
+            agentStore.sendOnboardingProgress(session.value.agent_run, {
+                flow_uuid: session.value.flow,
+                session_uuid: session.value.uuid,
+                page_uuid: currentPage.value.uuid,
+                page_order: currentPageIndex.value,
+                responses: normalizedResponses,
+                status: session.value.status,
+            })
+        }
+
+        if (hasNext.value) {
+            currentPageIndex.value += 1
+            hydrateFormState()
+        } else {
+            message.success('Onboarding completed')
+        }
+    } catch (err: unknown) {
+        console.error('Failed to submit onboarding page', err)
+        message.error('Failed to save onboarding progress')
+    }
+}
+
+const requestFeedback = async () => {
+    if (!currentPage.value || !session.value) return
+    feedbackLoading.value = true
+    try {
+        const normalizedResponses = normalizeResponses(formState.value)
+        const response = await apiClient.post<{ feedback: OnboardingFeedback; session: OnboardingSession }>(
+            API.onboardingSessionFeedback(session.value.uuid),
+            {
+                page_uuid: currentPage.value.uuid,
+                responses: normalizedResponses,
+            },
+        )
+        session.value = response.data.session
+        feedbackByPage.value[currentPage.value.uuid] = response.data.feedback
+    } catch (err: unknown) {
+        console.error('Failed to get feedback', err)
+        message.error('Failed to get feedback')
+    } finally {
+        feedbackLoading.value = false
+    }
+}
+
+const generateWithAgent = async () => {
+    if (!flowDetails.value) return
+    generating.value = true
+    try {
+        const response = await apiClient.post<OnboardingFlow>(
+            API.onboardingFlowGenerate(flowDetails.value.uuid),
+            {
+                instructions: generateInstructions.value,
+            },
+        )
+        flowDetails.value = response.data
+        currentPageIndex.value = 0
+        hydrateFormState()
+        message.success('Generated onboarding content')
+    } catch (err: unknown) {
+        console.error('Failed to generate onboarding flow', err)
+        message.error('Failed to generate onboarding content')
+    } finally {
+        generating.value = false
+    }
+}
+
+const publishFlow = async () => {
+    if (!flowDetails.value) return
+    publishing.value = true
+    try {
+        const response = await apiClient.post<OnboardingFlow>(
+            API.onboardingFlowPublish(flowDetails.value.uuid),
+        )
+        flowDetails.value = response.data
+        flows.value = flows.value.map((flow) =>
+            flow.uuid === response.data.uuid ? { ...flow, status: response.data.status } : flow,
+        )
+        message.success('Flow published')
+    } catch (err: unknown) {
+        console.error('Failed to publish onboarding flow', err)
+        message.error('Failed to publish onboarding flow')
+    } finally {
+        publishing.value = false
+    }
+}
+
+const createFlow = async () => {
+    if (!roleId.value) {
+        message.error('Role is required to create a flow')
+        return
+    }
+    if (!createFlowForm.value.title.trim()) {
+        message.error('Title is required')
+        return
+    }
+
+    creatingFlow.value = true
+    try {
+        const payload = {
+            role: roleId.value,
+            agent:
+                createFlowForm.value.agent !== undefined && createFlowForm.value.agent !== null
+                    ? String(createFlowForm.value.agent)
+                    : null,
+            title: createFlowForm.value.title.trim(),
+            description: createFlowForm.value.description.trim(),
+            status: 'draft',
+        }
+        const response = await apiClient.post<OnboardingFlow>(API.onboardingFlows(), payload)
+        flows.value = [response.data, ...flows.value]
+        selectedFlowUuid.value = response.data.uuid
+        createFlowForm.value = {
+            title: '',
+            description: '',
+            agent: undefined,
+        }
+        message.success('Onboarding flow created')
+    } catch (err: unknown) {
+        console.error('Failed to create onboarding flow', err)
+        message.error('Failed to create onboarding flow')
+    } finally {
+        creatingFlow.value = false
+    }
+}
+
+const goPrev = () => {
+    if (hasPrev.value) {
+        currentPageIndex.value -= 1
+        hydrateFormState()
+    }
+}
+
+const resetSession = async () => {
+    if (!flowDetails.value) return
+    resettingSession.value = true
+    try {
+        const response = await apiClient.post<OnboardingSession>(API.onboardingSessions(), {
+            flow: flowDetails.value.uuid,
+        })
+        session.value = response.data
+        currentPageIndex.value = 0
+        feedbackByPage.value = {}
+        hydrateFormState()
+        message.success('Started a new onboarding session')
+    } catch (err: unknown) {
+        console.error('Failed to reset onboarding session', err)
+        message.error('Failed to start a new onboarding session')
+    } finally {
+        resettingSession.value = false
+    }
+}
+
+watch(selectedFlowUuid, async (value) => {
+    if (value) {
+        await fetchFlowDetails(String(value))
+    }
+})
+
+watch(roleId, async () => {
+    selectedFlowUuid.value = undefined
+    flowDetails.value = null
+    session.value = null
+    userRoles.value = []
+    if (!roleId.value) {
+        await fetchUserRoles()
+        return
+    }
+    await fetchFlows()
+    if (selectedFlowUuid.value) {
+        await fetchFlowDetails(String(selectedFlowUuid.value))
+    }
+})
+
+watch(currentPageIndex, hydrateFormState)
+
+onUnmounted(() => {
+    agentStore.disconnect()
+})
+
+onMounted(async () => {
+    await userStore.fetchSession()
+    if (isManager.value) {
+        await fetchAgents()
+    }
+    if (!roleId.value) {
+        await fetchUserRoles()
+        return
+    }
+    await fetchFlows()
+    if (selectedFlowUuid.value) {
+        await fetchFlowDetails(String(selectedFlowUuid.value))
+    }
+})
+</script>
+
+<template>
+    <div class="page">
+        <Spin :spinning="loading" tip="Loading onboarding...">
+            <Card class="panel" :bordered="false">
+                <div class="header">
+                    <Typography.Title :level="2">Onboarding</Typography.Title>
+                    <Space v-if="flowDetails && isManager">
+                        <Input
+                            v-model:value="generateInstructions"
+                            placeholder="Optional generation instructions"
+                            style="width: 320px"
+                        />
+                        <Button
+                            type="primary"
+                            :loading="generating"
+                            :disabled="!flowDetails?.agent"
+                            @click="generateWithAgent"
+                        >
+                            Generate with Agent
+                        </Button>
+                        <Button
+                            :loading="publishing"
+                            :disabled="flowDetails?.status === 'published'"
+                            @click="publishFlow"
+                        >
+                            {{ flowDetails?.status === 'published' ? 'Published' : 'Publish' }}
+                        </Button>
+                    </Space>
+                </div>
+
+                <Typography.Paragraph v-if="loadError" type="danger">{{ loadError }}</Typography.Paragraph>
+
+                <div v-if="!hasRoleContext">
+                    <Typography.Paragraph type="secondary">
+                        Choose a role to continue onboarding.
+                    </Typography.Paragraph>
+                    <Typography.Paragraph v-if="roleLoadError" type="danger">
+                        {{ roleLoadError }}
+                    </Typography.Paragraph>
+                    <Spin :spinning="roleLoading">
+                        <div v-if="userRoles.length > 0" class="role-list">
+                            <List :data-source="userRoles" :bordered="false">
+                                <template #renderItem="{ item }">
+                                    <List.Item class="role-item">
+                                        <List.Item.Meta
+                                            :title="item.name"
+                                            :description="item.organization?.name || 'Organization'"
+                                        />
+                                        <Button type="primary" @click="router.push(`/onboarding/${item.uuid}`)">
+                                            Open
+                                        </Button>
+                                    </List.Item>
+                                </template>
+                            </List>
+                        </div>
+                        <Typography.Paragraph v-else type="secondary">
+                            You are not a member of any roles yet.
+                        </Typography.Paragraph>
+                    </Spin>
+                </div>
+
+                <div v-else-if="flows.length > 1" class="flow-select">
+                    <Typography.Text strong>Select a flow</Typography.Text>
+                    <Select
+                        :value="selectedFlowUuid"
+                        @update:value="(val) => (selectedFlowUuid = toSingleValue(val))"
+                        placeholder="Choose a flow"
+                        style="width: 320px"
+                        :options="flows.map((f) => ({ value: f.uuid, label: f.title }))"
+                    />
+                </div>
+
+                <div v-else-if="flows.length === 0">
+                    <Typography.Paragraph type="secondary">
+                        No onboarding flows available yet.
+                    </Typography.Paragraph>
+                    <div v-if="isManager" class="create-flow">
+                        <Typography.Text strong>Create a flow</Typography.Text>
+                        <Typography.Paragraph type="secondary">
+                            Create a draft flow and then generate pages with an agent.
+                        </Typography.Paragraph>
+                        <Typography.Paragraph v-if="!roleId" type="warning">
+                            Open this page with a role id to create a flow.
+                        </Typography.Paragraph>
+                        <Form layout="vertical" :model="createFlowForm" @finish="createFlow">
+                            <Form.Item
+                                label="Title"
+                                name="title"
+                                :rules="[{ required: true, message: 'Title is required' }]"
+                            >
+                                <Input v-model:value="createFlowForm.title" placeholder="Flow title" />
+                            </Form.Item>
+                            <Form.Item label="Description" name="description">
+                                <Input
+                                    v-model:value="createFlowForm.description"
+                                    placeholder="Short description"
+                                />
+                            </Form.Item>
+                            <Form.Item label="Agent (optional)" name="agent">
+                                <Select
+                                    :value="createFlowForm.agent"
+                                    @update:value="(val) => (createFlowForm.agent = toSingleValue(val))"
+                                    allow-clear
+                                    placeholder="Select an agent"
+                                    :options="agents.map((a) => ({ value: a.uuid, label: a.name }))"
+                                />
+                            </Form.Item>
+                            <Button
+                                type="primary"
+                                html-type="submit"
+                                :loading="creatingFlow"
+                                :disabled="!roleId"
+                            >
+                                Create draft flow
+                            </Button>
+                        </Form>
+                        <Typography.Paragraph v-if="!agents.length" type="secondary">
+                            No agents available yet. You can create a flow now and assign an agent later.
+                        </Typography.Paragraph>
+                    </div>
+                </div>
+
+                <div v-if="flowDetails" class="flow-body">
+                    <Typography.Title :level="3">{{ flowDetails.title }}</Typography.Title>
+                    <Typography.Paragraph type="secondary">
+                        {{ flowDetails.description || 'No description provided.' }}
+                    </Typography.Paragraph>
+                    <Space v-if="session" class="session-actions">
+                        <Button danger :loading="resettingSession" @click="resetSession">
+                            Start New Session
+                        </Button>
+                    </Space>
+
+                    <Divider />
+
+                    <div v-if="currentPage" class="page-body">
+                        <Typography.Title :level="4">
+                            {{ currentPage.title }}
+                        </Typography.Title>
+                        <Typography.Paragraph>
+                            {{ currentPage.body }}
+                        </Typography.Paragraph>
+
+                        <Form layout="vertical" :model="formState" @finish="onSubmitPage">
+                            <Form.Item
+                                v-for="field in currentPage.fields"
+                                :key="field.uuid"
+                                :label="field.label"
+                                :name="field.key"
+                                :rules="field.required ? [{ required: true, message: 'Required' }] : []"
+                            >
+                                <Input
+                                    v-if="field.field_type === 'text'"
+                                    :value="(getFieldValue(field.key) as string | number | undefined)"
+                                    @update:value="(val) => setFieldValue(field.key, val)"
+                                    :placeholder="field.placeholder"
+                                />
+                                <Input.TextArea
+                                    v-else-if="field.field_type === 'textarea'"
+                                    :value="(getFieldValue(field.key) as string | undefined)"
+                                    @update:value="(val) => setFieldValue(field.key, val)"
+                                    :placeholder="field.placeholder"
+                                    :rows="4"
+                                />
+                                <InputNumber
+                                    v-else-if="field.field_type === 'number'"
+                                    :value="toNumberValue(getFieldValue(field.key))"
+                                    @update:value="(val) => setFieldValue(field.key, val)"
+                                    style="width: 100%"
+                                />
+                                <Select
+                                    v-else-if="field.field_type === 'select'"
+                                    :value="(getFieldValue(field.key) as string | undefined)"
+                                    @update:value="(val) => setFieldValue(field.key, val)"
+                                    :options="(field.options || []).map((opt) => ({ value: opt, label: opt }))"
+                                />
+                                <Select
+                                    v-else-if="field.field_type === 'multiselect'"
+                                    :value="(getFieldValue(field.key) as string[] | undefined)"
+                                    @update:value="(val) => setFieldValue(field.key, val)"
+                                    mode="multiple"
+                                    :options="(field.options || []).map((opt) => ({ value: opt, label: opt }))"
+                                />
+                                <Switch
+                                    v-else-if="field.field_type === 'boolean'"
+                                    :checked="Boolean(getFieldValue(field.key))"
+                                    @update:checked="(val) => setFieldValue(field.key, val)"
+                                />
+                                <DatePicker
+                                    v-else-if="field.field_type === 'date'"
+                                    :value="toDateValue(getFieldValue(field.key))"
+                                    @update:value="(val) => setFieldValue(field.key, val)"
+                                    style="width: 100%"
+                                />
+                                <Input
+                                    v-else
+                                    :value="(getFieldValue(field.key) as string | number | undefined)"
+                                    @update:value="(val) => setFieldValue(field.key, val)"
+                                    :placeholder="field.placeholder"
+                                />
+                            </Form.Item>
+
+                            <Space>
+                                <Button :disabled="!hasPrev" @click="goPrev">Back</Button>
+                                <Button :loading="feedbackLoading" @click="requestFeedback">
+                                    Get Feedback
+                                </Button>
+                                <Button type="primary" html-type="submit">
+                                    {{ hasNext ? 'Next' : 'Finish' }}
+                                </Button>
+                            </Space>
+                            <div class="feedback">
+                                <div v-if="feedbackByPage[currentPage.uuid]" class="feedback-panel">
+                                    <Typography.Text strong>Feedback</Typography.Text>
+                                    <Typography.Paragraph>
+                                        {{ feedbackByPage[currentPage.uuid]?.summary || '' }}
+                                    </Typography.Paragraph>
+                                </div>
+                            </div>
+                        </Form>
+                    </div>
+                </div>
+            </Card>
+        </Spin>
+    </div>
+</template>
+
+<style scoped>
+.page {
+    max-width: 1100px;
+    margin: 0 auto;
+    padding: 1rem;
+}
+.panel {
+    background: #0f172a;
+    border: 1px solid #1f2937;
+    color: #e5e7eb;
+}
+.header {
+    display: flex;
+    justify-content: space-between;
+    align-items: center;
+    gap: 1rem;
+    margin-bottom: 1rem;
+}
+.flow-select {
+    margin: 1rem 0;
+}
+.role-list {
+    margin: 1rem 0;
+}
+.role-item :deep(.ant-list-item-meta-title),
+.role-item :deep(.ant-list-item-meta-description) {
+    background: #0f172a;
+    border: 1px solid #1f2937;
+}
+.flow-body {
+    margin-top: 1rem;
+}
+.session-actions {
+    margin-top: 0.5rem;
+}
+.create-flow {
+    margin-top: 1rem;
+    padding: 1rem;
+    border: 1px dashed #1f2937;
+    border-radius: 6px;
+    background: #0b1220;
+}
+.page-body {
+    margin-top: 1rem;
+}
+.feedback {
+    margin-top: 1rem;
+    display: grid;
+    gap: 0.75rem;
+}
+.feedback-panel {
+    padding: 0.75rem 1rem;
+    border: 1px solid #1f2937;
+    border-radius: 6px;
+    background: #0b1220;
+}
+.feedback-panel ul {
+    margin: 0.5rem 0 0.75rem 1.25rem;
+}
+</style>

src/views/OrganizationView.vue

@@ -1,7 +1,7 @@
 <script setup lang="ts">
 import { ref, onMounted, computed, h } from 'vue'
 import { useRouter, useRoute } from 'vue-router'
-import { Card, Typography, Button, List, Space, Spin, message, Tag, Divider, Upload, Modal, Table } from 'ant-design-vue'
+import { Card, Typography, Button, List, Space, Spin, message, Tag, Divider, Upload, Modal, Table, Select } from 'ant-design-vue'
 import { apiClient, isAxiosError, API } from '../router/api'
 import { useUserStore } from '../stores/userStore'
 import { InboxOutlined, DeleteOutlined } from '@ant-design/icons-vue'
@@ -52,28 +52,13 @@ const fetchRoles = async () => {
 }
 
 const fetchUserRoleMemberships = async () => {
-    const userRoleUuids: string[] = []
+    if (!organization.value?.uuid) return
-    const userId = auth.user?.id
-    if (!userId || !organization.value?.uuid) return
     try {
-        const checks = roles.value.map(async (r) => {
+        const response = await apiClient.get<Role[]>(API.organizationRolesMine())
-            try {
+        const mine = Array.isArray(response.data) ? response.data : []
-                const resp = await apiClient.get<Array<{ user: { id: number } }>>(
+        const orgUuid = organization.value.uuid
-                    API.organizationRoleMembers(organization.value!.uuid, r.uuid),
+        const joinedRoles = mine.filter((role) => role.organization?.uuid === orgUuid)
-                )
+        auth.setJoinedRoles(joinedRoles)
-                const found =
-                    Array.isArray(resp.data) && resp.data.some((m) => m.user?.id === userId)
-                if (found && r.uuid) userRoleUuids.push(r.uuid)
-            } catch {
-                // ignore individual role errors
-            }
-        })
-        await Promise.all(checks)
-        // update the global store with actual Role objects the user has joined
-        const joinedRoles = roles.value.filter((r) => userRoleUuids.includes(r.uuid))
-        if (joinedRoles.length) {
-            auth.setJoinedRoles(joinedRoles)
-        }
     } catch (err) {
         console.error('Failed to fetch user role memberships', err)
     }
@@ -169,6 +154,7 @@ const beforeUpload = (file: File) => {
 
 const selectedFile = ref<File | null>(null)
 const fileDescription = ref('')
+const selectedRoleUuid = ref<string>('')
 
 const handleFileSelected = (file: File) => {
     selectedFile.value = file
@@ -179,10 +165,15 @@ const handleFileUploadClick = async () => {
         message.error('Please select a file to upload')
         return
     }
+    if (!selectedRoleUuid.value) {
+        message.error('Please select a role for this training file')
+        return
+    }
 
     await handleFileUpload(selectedFile.value, fileDescription.value)
     selectedFile.value = null
     fileDescription.value = ''
+    selectedRoleUuid.value = ''
 }
 
 const handleFileUpload = async (file: File, description: string = '') => {
@@ -197,6 +188,9 @@ const handleFileUpload = async (file: File, description: string = '') => {
         formData.append('file', file)
         formData.append('file_name', file.name)
         formData.append('description', description)
+        if (selectedRoleUuid.value) {
+            formData.append('role_uuid', selectedRoleUuid.value)
+        }
 
         const response = await apiClient.post<TrainingFile>(
             API.organizationTrainingFiles(organization.value.uuid),
@@ -278,12 +272,25 @@ const trainingFileColumns = [
         key: 'file_size',
         customRender: ({ value }: { value: number }) => formatFileSize(value || 0),
     },
+    {
+        title: 'Role',
+        key: 'role',
+        customRender: ({ record }: { record: TrainingFile }) => record.role?.name || '-',
+    },
     {
         title: 'Status',
-        dataIndex: 'is_processed',
+        dataIndex: 'status',
-        key: 'is_processed',
+        key: 'status',
-        customRender: ({ value }: { value: boolean }) =>
+        customRender: ({ value }: { value: string }) => {
-            h(Tag, { color: value ? 'success' : 'processing' }, () => value ? 'Processed' : 'Processing'),
+            const statusMap: Record<string, { color: string; label: string }> = {
+                ingesting: { color: 'processing', label: 'Ingesting' },
+                chunked: { color: 'blue', label: 'Chunked' },
+                embedded: { color: 'success', label: 'Embedded' },
+                failed: { color: 'error', label: 'Failed' },
+            }
+            const status = statusMap[value] || { color: 'default', label: value }
+            return h(Tag, { color: status.color }, () => status.label)
+        },
     },
     {
         title: 'Uploaded',
@@ -312,13 +319,13 @@ const trainingFileColumns = [
     },
 ]
 
-onMounted(() => {
+onMounted(async () => {
-    fetchOrganization().then(async () => {
+    await auth.fetchSession(true)
-        await fetchMembers()
+    await fetchOrganization()
-        await fetchRoles()
+    await fetchMembers()
-        await fetchUserRoleMemberships()
+    await fetchRoles()
-        await fetchTrainingFiles()
+    await fetchUserRoleMemberships()
-    })
+    await fetchTrainingFiles()
 })
 </script>
 
@@ -404,7 +411,6 @@ onMounted(() => {
                                     <Button
                                         type="default"
                                         size="small"
-                                        :disabled="isManager"
                                         @click="router.push(`/onboarding/${item.uuid}`)"
                                     >
                                         Start Onboarding
@@ -418,7 +424,9 @@ onMounted(() => {
                                     >
                                         Join Role
                                     </Button>
-                                    <Tag v-else color="success">Joined</Tag>
+                                    <Button v-else size="small" disabled>
+                                        Joined
+                                    </Button>
                                 </Space>
                             </List.Item>
                         </template>
@@ -436,7 +444,7 @@ onMounted(() => {
             width="600px"
             ok-text="Upload"
             cancel-text="Cancel"
-            :ok-button-props="{ loading: uploading, disabled: !selectedFile }"
+            :ok-button-props="{ loading: uploading, disabled: !selectedFile || !selectedRoleUuid }"
             @ok="handleFileUploadClick"
             @cancel="showUploadModal = false"
         >
@@ -444,6 +452,15 @@ onMounted(() => {
                 <Typography.Text>
                     Supported formats: <strong>txt, pdf, md, csv, json, docx, doc</strong> (Max 50MB)
                 </Typography.Text>
+                <div>
+                    <Typography.Text strong>Role</Typography.Text>
+                    <Select
+                        v-model:value="selectedRoleUuid"
+                        placeholder="Select a role"
+                        style="width: 100%"
+                        :options="roles.map((role) => ({ label: role.name, value: role.uuid }))"
+                    />
+                </div>
                 <Upload.Dragger
                     accept=".txt,.pdf,.md,.csv,.json,.docx,.doc"
                     :before-upload="(file) => {