new feature

Reviewed-on: #15

.env.example

@@ -27,6 +27,9 @@ CHROMADB_PATH=./data/chromadb
 # OpenAI Configuration
 OPENAI_API_KEY=your-openai-api-key
 
+# Tavily Configuration (for web search)
+TAVILY_API_KEY=your-tavily-api-key
+
 # Immich Configuration
 IMMICH_URL=http://192.168.1.5:2283
 IMMICH_API_KEY=your-immich-api-key
@@ -45,3 +48,38 @@ OIDC_USE_DISCOVERY=true
 # OIDC_TOKEN_ENDPOINT=https://auth.example.com/api/oidc/token
 # OIDC_USERINFO_ENDPOINT=https://auth.example.com/api/oidc/userinfo
 # OIDC_JWKS_URI=https://auth.example.com/api/oidc/jwks
+
+# YNAB Configuration
+# Get your Personal Access Token from https://app.ynab.com/settings/developer
+YNAB_ACCESS_TOKEN=your-ynab-personal-access-token
+# Optional: Specify a budget ID, or leave empty to use the default/first budget
+YNAB_BUDGET_ID=
+
+# Twilio Configuration (WhatsApp)
+TWILIO_ACCOUNT_SID=your-twilio-account-sid
+TWILIO_AUTH_TOKEN=your-twilio-auth-token
+TWILIO_WHATSAPP_NUMBER=whatsapp:+14155238886
+# Comma-separated list of WhatsApp numbers allowed to use the service (e.g., whatsapp:+1234567890)
+# Use * to allow any number
+ALLOWED_WHATSAPP_NUMBERS=
+# Set to false to disable Twilio signature validation in development
+TWILIO_SIGNATURE_VALIDATION=true
+# If behind a reverse proxy, set this to your public webhook URL so signature validation works
+# TWILIO_WEBHOOK_URL=https://your-domain.com/api/whatsapp/webhook
+# Rate limiting: max messages per window (default: 10 messages per 60 seconds)
+# WHATSAPP_RATE_LIMIT_MAX=10
+# WHATSAPP_RATE_LIMIT_WINDOW=60
+
+# Obsidian Configuration (headless sync)
+# Auth token from Obsidian account (Settings → Account → API token)
+OBSIDIAN_AUTH_TOKEN=your-obsidian-auth-token
+# Vault ID to sync (found in Obsidian sync settings)
+OBSIDIAN_VAULT_ID=your-vault-id
+# End-to-end encryption password (if vault uses E2E encryption)
+OBSIDIAN_E2E_PASSWORD=
+# Device name shown in Obsidian sync activity
+OBSIDIAN_DEVICE_NAME=simbarag
+# Set to true to run continuous sync in the background
+OBSIDIAN_CONTINUOUS_SYNC=false
+# Local path to Obsidian vault (where files are synced)
+OBSIDIAN_VAULT_PATH=/app/data/obsidian

.envrc

@@ -0,0 +1 @@
+dotenv_if_exists

.gitignore

@@ -18,3 +18,6 @@ chromadb_openai/
 chroma_db/
 database/
 *.db
+
+obvault/
+.claude

CLAUDE.md

@@ -11,21 +11,21 @@ SimbaRAG is a RAG (Retrieval-Augmented Generation) conversational AI system for
 ### Development
 
 ```bash
-# Start dev environment with hot reload
+# Start environment
-docker compose -f docker-compose.dev.yml up --build
+docker compose up --build
 
 # View logs
-docker compose -f docker-compose.dev.yml logs -f raggr
+docker compose logs -f raggr
 ```
 
 ### Database Migrations (Aerich/Tortoise ORM)
 
 ```bash
 # Generate migration (must run in Docker with DB access)
-docker compose -f docker-compose.dev.yml exec raggr aerich migrate --name describe_change
+docker compose exec raggr aerich migrate --name describe_change
 
 # Apply migrations (auto-runs on startup, manual if needed)
-docker compose -f docker-compose.dev.yml exec raggr aerich upgrade
+docker compose exec raggr aerich upgrade
 
 # View migration history
 docker compose exec raggr aerich history

Dockerfile

@@ -6,9 +6,9 @@ WORKDIR /app
 RUN apt-get update && apt-get install -y \
     build-essential \
     curl \
-    && curl -fsSL https://deb.nodesource.com/setup_20.x | bash - \
+    && curl -fsSL https://deb.nodesource.com/setup_22.x | bash - \
     && apt-get install -y nodejs \
-    && npm install -g yarn \
+    && npm install -g yarn obsidian-headless \
     && rm -rf /var/lib/apt/lists/* \
     && curl -LsSf https://astral.sh/uv/install.sh | sh
 

aerich_config.py

@@ -1,4 +1,8 @@
 import os
+from dotenv import load_dotenv
+
+# Load environment variables
+load_dotenv()
 
 # Database configuration with environment variable support
 # Use DATABASE_PATH for relative paths or DATABASE_URL for full connection strings

app.py

@@ -1,16 +1,25 @@
+import logging
 import os
 
+from dotenv import load_dotenv
 from quart import Quart, jsonify, render_template, request, send_from_directory
 from quart_jwt_extended import JWTManager, get_jwt_identity, jwt_refresh_token_required
-from tortoise.contrib.quart import register_tortoise
+from tortoise import Tortoise
 
 import blueprints.conversation
 import blueprints.conversation.logic
 import blueprints.rag
 import blueprints.users
+import blueprints.whatsapp
 import blueprints.users.models
 from main import consult_simba_oracle
 
+# Load environment variables
+load_dotenv()
+
+# Configure logging
+logging.basicConfig(level=logging.INFO)
+
 app = Quart(
     __name__,
     static_folder="raggr-frontend/dist/static",
@@ -24,6 +33,7 @@ jwt = JWTManager(app)
 app.register_blueprint(blueprints.users.user_blueprint)
 app.register_blueprint(blueprints.conversation.conversation_blueprint)
 app.register_blueprint(blueprints.rag.rag_blueprint)
+app.register_blueprint(blueprints.whatsapp.whatsapp_blueprint)
 
 
 # Database configuration with environment variable support
@@ -44,12 +54,15 @@ TORTOISE_CONFIG = {
     },
 }
 
-# Initialize Tortoise ORM
+# Initialize Tortoise ORM with lifecycle hooks
-register_tortoise(
+@app.while_serving
-    app,
+async def lifespan():
-    config=TORTOISE_CONFIG,
+    logging.info("Initializing Tortoise ORM...")
-    generate_schemas=False,  # Disabled - using Aerich for migrations
+    await Tortoise.init(config=TORTOISE_CONFIG)
-)
+    logging.info("Tortoise ORM initialized successfully")
+    yield
+    logging.info("Closing Tortoise ORM connections...")
+    await Tortoise.close_connections()
 
 
 # Serve React static files

blueprints/conversation/__init__.py

@@ -1,6 +1,7 @@
 import datetime
+import json
 
-from quart import Blueprint, jsonify, request
+from quart import Blueprint, jsonify, make_response, request
 from quart_jwt_extended import (
     get_jwt_identity,
     jwt_refresh_token_required,
@@ -24,35 +25,7 @@ conversation_blueprint = Blueprint(
     "conversation_api", __name__, url_prefix="/api/conversation"
 )
 
-
+_SYSTEM_PROMPT = """You are a helpful cat assistant named Simba that understands veterinary terms. When there are questions to you specifically, they are referring to Simba the cat. Answer the user in as if you were a cat named Simba. Don't act too catlike. Be assertive.
-@conversation_blueprint.post("/query")
-@jwt_refresh_token_required
-async def query():
-    current_user_uuid = get_jwt_identity()
-    user = await blueprints.users.models.User.get(id=current_user_uuid)
-    data = await request.get_json()
-    query = data.get("query")
-    conversation_id = data.get("conversation_id")
-    conversation = await get_conversation_by_id(conversation_id)
-    await conversation.fetch_related("messages")
-    await add_message_to_conversation(
-        conversation=conversation,
-        message=query,
-        speaker="user",
-        user=user,
-    )
-
-    # Build conversation history from recent messages (last 10 for context)
-    recent_messages = (
-        conversation.messages[-10:]
-        if len(conversation.messages) > 10
-        else conversation.messages
-    )
-
-    messages_payload = [
-        {
-            "role": "system",
-            "content": """You are a helpful cat assistant named Simba that understands veterinary terms. When there are questions to you specifically, they are referring to Simba the cat. Answer the user in as if you were a cat named Simba. Don't act too catlike. Be assertive.
 
 SIMBA FACTS (as of January 2026):
 - Name: Simba
@@ -84,18 +57,65 @@ Upcoming Appointments:
 - Routine Examination: Due 6/1/2026
 - FVRCP-3yr Vaccine: Due 10/2/2026
 
-IMPORTANT: When users ask factual questions about Simba's health, medical history, veterinary visits, medications, weight, or any information that would be in documents, you MUST use the simba_search tool to retrieve accurate information before answering. Do not rely on general knowledge - always search the documents for factual questions.""",
+IMPORTANT: When users ask factual questions about Simba's health, medical history, veterinary visits, medications, weight, or any information that would be in documents, you MUST use the simba_search tool to retrieve accurate information before answering. Do not rely on general knowledge - always search the documents for factual questions.
-        }
-    ]
 
-    # Add recent conversation history
+BUDGET & FINANCE (YNAB Integration):
+You have access to Ryan's budget data through YNAB (You Need A Budget). When users ask about financial matters, use the appropriate YNAB tools:
+- Use ynab_budget_summary for overall budget health and status questions
+- Use ynab_search_transactions to find specific purchases or spending at particular stores
+- Use ynab_category_spending to analyze spending by category for a month
+- Use ynab_insights to provide spending trends, patterns, and recommendations
+Always use these tools when asked about budgets, spending, transactions, or financial health.
+
+NOTES & RESEARCH (Obsidian Integration):
+You have access to Ryan's Obsidian vault through the Obsidian integration. When users ask about research, personal notes, or information that might be stored in markdown files, use the appropriate Obsidian tools:
+- Use obsidian_search_notes to search through your vault for relevant information
+- Use obsidian_read_note to read the full content of a specific note by path
+- Use obsidian_create_note to save new findings, ideas, or research to your vault
+- Use obsidian_create_task to create task notes with due dates
+Always use these tools when users ask about notes, research, ideas, tasks, or when you want to save information for future reference.
+
+DAILY JOURNAL (Task Tracking):
+You have access to Ryan's daily journal notes. Each note lives at journal/YYYY/YYYY-MM-DD.md and has two sections: tasks and log.
+- Use journal_get_today to read today's full daily note (tasks + log)
+- Use journal_get_tasks to list tasks (done/pending) for today or a specific date
+- Use journal_add_task to add a new task to today's (or a given date's) note
+- Use journal_complete_task to check off a task as done
+Use these tools when Ryan asks about today's tasks, wants to add something to his list, or wants to mark a task complete."""
+
+
+def _build_messages_payload(conversation, query_text: str) -> list:
+    recent_messages = (
+        conversation.messages[-10:]
+        if len(conversation.messages) > 10
+        else conversation.messages
+    )
+    messages_payload = [{"role": "system", "content": _SYSTEM_PROMPT}]
     for msg in recent_messages[:-1]:  # Exclude the message we just added
         role = "user" if msg.speaker == "user" else "assistant"
         messages_payload.append({"role": role, "content": msg.text})
+    messages_payload.append({"role": "user", "content": query_text})
+    return messages_payload
 
-    # Add current query
-    messages_payload.append({"role": "user", "content": query})
 
+@conversation_blueprint.post("/query")
+@jwt_refresh_token_required
+async def query():
+    current_user_uuid = get_jwt_identity()
+    user = await blueprints.users.models.User.get(id=current_user_uuid)
+    data = await request.get_json()
+    query = data.get("query")
+    conversation_id = data.get("conversation_id")
+    conversation = await get_conversation_by_id(conversation_id)
+    await conversation.fetch_related("messages")
+    await add_message_to_conversation(
+        conversation=conversation,
+        message=query,
+        speaker="user",
+        user=user,
+    )
+
+    messages_payload = _build_messages_payload(conversation, query)
     payload = {"messages": messages_payload}
 
     response = await main_agent.ainvoke(payload)
@@ -109,6 +129,75 @@ IMPORTANT: When users ask factual questions about Simba's health, medical histor
     return jsonify({"response": message})
 
 
+@conversation_blueprint.post("/stream-query")
+@jwt_refresh_token_required
+async def stream_query():
+    current_user_uuid = get_jwt_identity()
+    user = await blueprints.users.models.User.get(id=current_user_uuid)
+    data = await request.get_json()
+    query_text = data.get("query")
+    conversation_id = data.get("conversation_id")
+    conversation = await get_conversation_by_id(conversation_id)
+    await conversation.fetch_related("messages")
+    await add_message_to_conversation(
+        conversation=conversation,
+        message=query_text,
+        speaker="user",
+        user=user,
+    )
+
+    messages_payload = _build_messages_payload(conversation, query_text)
+    payload = {"messages": messages_payload}
+
+    async def event_generator():
+        final_message = None
+        try:
+            async for event in main_agent.astream_events(payload, version="v2"):
+                event_type = event.get("event")
+
+                if event_type == "on_tool_start":
+                    yield f"data: {json.dumps({'type': 'tool_start', 'tool': event['name']})}\n\n"
+
+                elif event_type == "on_tool_end":
+                    yield f"data: {json.dumps({'type': 'tool_end', 'tool': event['name']})}\n\n"
+
+                elif event_type == "on_chain_end":
+                    output = event.get("data", {}).get("output")
+                    if isinstance(output, dict):
+                        msgs = output.get("messages", [])
+                        if msgs:
+                            last_msg = msgs[-1]
+                            content = getattr(last_msg, "content", None)
+                            if isinstance(content, str) and content:
+                                final_message = content
+
+        except Exception as e:
+            yield f"data: {json.dumps({'type': 'error', 'message': str(e)})}\n\n"
+
+        if final_message:
+            await add_message_to_conversation(
+                conversation=conversation,
+                message=final_message,
+                speaker="simba",
+                user=user,
+            )
+            yield f"data: {json.dumps({'type': 'response', 'message': final_message})}\n\n"
+        else:
+            yield f"data: {json.dumps({'type': 'error', 'message': 'No response generated'})}\n\n"
+
+        yield "data: [DONE]\n\n"
+
+    return await make_response(
+        event_generator(),
+        200,
+        {
+            "Content-Type": "text/event-stream",
+            "Cache-Control": "no-cache",
+            "X-Accel-Buffering": "no",
+        },
+    )
+
+
 @conversation_blueprint.route("/<conversation_id>")
 @jwt_refresh_token_required
 async def get_conversation(conversation_id: str):

blueprints/conversation/agents.py

@@ -1,6 +1,7 @@
 import os
 from typing import cast
 
+from dotenv import load_dotenv
 from langchain.agents import create_agent
 from langchain.chat_models import BaseChatModel
 from langchain.tools import tool
@@ -8,6 +9,11 @@ from langchain_openai import ChatOpenAI
 from tavily import AsyncTavilyClient
 
 from blueprints.rag.logic import query_vector_store
+from utils.obsidian_service import ObsidianService
+from utils.ynab_service import YNABService
+
+# Load environment variables
+load_dotenv()
 
 # Configure LLM with llama-server or OpenAI fallback
 llama_url = os.getenv("LLAMA_SERVER_URL")
@@ -25,7 +31,41 @@ model_with_fallback = cast(
     BaseChatModel,
     llama_chat.with_fallbacks([openai_fallback]) if llama_chat else openai_fallback,
 )
-client = AsyncTavilyClient(os.getenv("TAVILY_KEY"), "")
+client = AsyncTavilyClient(api_key=os.getenv("TAVILY_API_KEY", ""))
+
+# Initialize YNAB service (will only work if YNAB_ACCESS_TOKEN is set)
+try:
+    ynab_service = YNABService()
+    ynab_enabled = True
+except (ValueError, Exception) as e:
+    print(f"YNAB service not initialized: {e}")
+    ynab_enabled = False
+
+# Initialize Obsidian service (will only work if OBSIDIAN_VAULT_PATH is set)
+try:
+    obsidian_service = ObsidianService()
+    obsidian_enabled = True
+except (ValueError, Exception) as e:
+    print(f"Obsidian service not initialized: {e}")
+    obsidian_enabled = False
+
+
+@tool
+def get_current_date() -> str:
+    """Get today's date in a human-readable format.
+
+    Use this tool when you need to:
+    - Reference today's date in your response
+    - Answer questions like "what is today's date"
+    - Format dates in messages or documents
+    - Calculate time periods relative to today
+
+    Returns:
+        Today's date in YYYY-MM-DD format
+    """
+    from datetime import date
+
+    return date.today().isoformat()
 
 
 @tool
@@ -85,4 +125,494 @@ async def simba_search(query: str):
     return serialized, docs
 
 
-main_agent = create_agent(model=model_with_fallback, tools=[simba_search, web_search])
+@tool
+def ynab_budget_summary() -> str:
+    """Get overall budget summary and health status from YNAB.
+
+    Use this tool when the user asks about:
+    - Overall budget health or status
+    - How much money is to be budgeted
+    - Total budget amounts or spending
+    - General budget overview questions
+
+    Returns:
+        Summary of budget health, to-be-budgeted amount, total budgeted,
+        total activity, and available amounts.
+    """
+    if not ynab_enabled:
+        return "YNAB integration is not configured. Please set YNAB_ACCESS_TOKEN environment variable."
+
+    try:
+        summary = ynab_service.get_budget_summary()
+        return summary["summary"]
+    except Exception as e:
+        return f"Error fetching budget summary: {str(e)}"
+
+
+@tool
+def ynab_search_transactions(
+    start_date: str = "",
+    end_date: str = "",
+    category_name: str = "",
+    payee_name: str = "",
+) -> str:
+    """Search YNAB transactions by date range, category, or payee.
+
+    Use this tool when the user asks about:
+    - Specific transactions or purchases
+    - Spending at a particular store or payee
+    - Transactions in a specific category
+    - What was spent during a time period
+
+    Args:
+        start_date: Start date in YYYY-MM-DD format (optional, defaults to 30 days ago)
+        end_date: End date in YYYY-MM-DD format (optional, defaults to today)
+        category_name: Filter by category name (optional, partial match)
+        payee_name: Filter by payee/store name (optional, partial match)
+
+    Returns:
+        List of matching transactions with dates, amounts, categories, and payees.
+    """
+    if not ynab_enabled:
+        return "YNAB integration is not configured. Please set YNAB_ACCESS_TOKEN environment variable."
+
+    try:
+        result = ynab_service.get_transactions(
+            start_date=start_date or None,
+            end_date=end_date or None,
+            category_name=category_name or None,
+            payee_name=payee_name or None,
+        )
+
+        if result["count"] == 0:
+            return "No transactions found matching the specified criteria."
+
+        # Format transactions for readability
+        txn_list = []
+        for txn in result["transactions"][:10]:  # Limit to 10 for readability
+            txn_list.append(
+                f"- {txn['date']}: {txn['payee']} - ${abs(txn['amount']):.2f} ({txn['category'] or 'Uncategorized'})"
+            )
+
+        return (
+            f"Found {result['count']} transactions from {result['start_date']} to {result['end_date']}. "
+            f"Total: ${abs(result['total_amount']):.2f}\n\n"
+            + "\n".join(txn_list)
+            + (
+                f"\n\n(Showing first 10 of {result['count']} transactions)"
+                if result["count"] > 10
+                else ""
+            )
+        )
+    except Exception as e:
+        return f"Error searching transactions: {str(e)}"
+
+
+@tool
+def ynab_category_spending(month: str = "") -> str:
+    """Get spending breakdown by category for a specific month.
+
+    Use this tool when the user asks about:
+    - Spending by category
+    - What categories were overspent
+    - Monthly spending breakdown
+    - Budget vs actual spending for a month
+
+    Args:
+        month: Month in YYYY-MM format (optional, defaults to current month)
+
+    Returns:
+        Spending breakdown by category with budgeted, spent, and available amounts.
+    """
+    if not ynab_enabled:
+        return "YNAB integration is not configured. Please set YNAB_ACCESS_TOKEN environment variable."
+
+    try:
+        result = ynab_service.get_category_spending(month=month or None)
+
+        summary = (
+            f"Budget spending for {result['month']}:\n"
+            f"Total budgeted: ${result['total_budgeted']:.2f}\n"
+            f"Total spent: ${result['total_spent']:.2f}\n"
+            f"Total available: ${result['total_available']:.2f}\n"
+        )
+
+        if result["overspent_categories"]:
+            summary += (
+                f"\nOverspent categories ({len(result['overspent_categories'])}):\n"
+            )
+            for cat in result["overspent_categories"][:5]:
+                summary += f"- {cat['name']}: Budgeted ${cat['budgeted']:.2f}, Spent ${cat['spent']:.2f}, Over by ${cat['overspent_by']:.2f}\n"
+
+        # Add top spending categories
+        summary += "\nTop spending categories:\n"
+        for cat in result["categories"][:10]:
+            if cat["activity"] < 0:  # Only show spending (negative activity)
+                summary += f"- {cat['category']}: ${abs(cat['activity']):.2f} (budgeted: ${cat['budgeted']:.2f}, available: ${cat['available']:.2f})\n"
+
+        return summary
+    except Exception as e:
+        return f"Error fetching category spending: {str(e)}"
+
+
+@tool
+def ynab_insights(months_back: int = 3) -> str:
+    """Generate insights about spending patterns and budget health over time.
+
+    Use this tool when the user asks about:
+    - Spending trends or patterns
+    - Budget recommendations
+    - Which categories are frequently overspent
+    - How current spending compares to past months
+    - Overall budget health analysis
+
+    Args:
+        months_back: Number of months to analyze (default 3, max 6)
+
+    Returns:
+        Insights about spending trends, frequently overspent categories,
+        and personalized recommendations.
+    """
+    if not ynab_enabled:
+        return "YNAB integration is not configured. Please set YNAB_ACCESS_TOKEN environment variable."
+
+    try:
+        # Limit to reasonable range
+        months_back = min(max(1, months_back), 6)
+        result = ynab_service.get_spending_insights(months_back=months_back)
+
+        if "error" in result:
+            return result["error"]
+
+        summary = (
+            f"Spending insights for the last {months_back} months:\n\n"
+            f"Average monthly spending: ${result['average_monthly_spending']:.2f}\n"
+            f"Current month spending: ${result['current_month_spending']:.2f}\n"
+            f"Spending trend: {result['spending_trend']}\n"
+        )
+
+        if result["frequently_overspent_categories"]:
+            summary += "\nFrequently overspent categories:\n"
+            for cat in result["frequently_overspent_categories"][:5]:
+                summary += f"- {cat['category']}: overspent in {cat['months_overspent']} of {months_back} months\n"
+
+        if result["recommendations"]:
+            summary += "\nRecommendations:\n"
+            for rec in result["recommendations"]:
+                summary += f"- {rec}\n"
+
+        return summary
+    except Exception as e:
+        return f"Error generating insights: {str(e)}"
+
+
+@tool
+async def obsidian_search_notes(query: str) -> str:
+    """Search through Obsidian vault notes for information.
+
+    Use this tool when you need to:
+    - Find information in personal notes
+    - Research past ideas or thoughts from your vault
+    - Look up information stored in markdown files
+    - Search for content that would be in your notes
+
+    Args:
+        query: The search query to look up in your Obsidian vault
+
+    Returns:
+        Relevant notes with their content and metadata
+    """
+    if not obsidian_enabled:
+        return "Obsidian integration is not configured. Please set OBSIDIAN_VAULT_PATH environment variable."
+
+    try:
+        # Query ChromaDB for obsidian documents
+        serialized, docs = await query_vector_store(query=query)
+        return serialized
+
+    except Exception as e:
+        return f"Error searching Obsidian notes: {str(e)}"
+
+
+@tool
+async def obsidian_read_note(relative_path: str) -> str:
+    """Read a specific note from your Obsidian vault.
+
+    Use this tool when you want to:
+    - Read the full content of a specific note
+    - Get detailed information from a particular markdown file
+    - Access content from a known note path
+
+    Args:
+        relative_path: Path to note relative to vault root (e.g., "notes/my-note.md")
+
+    Returns:
+        Full content and metadata of the requested note
+    """
+    if not obsidian_enabled:
+        return "Obsidian integration is not configured. Please set OBSIDIAN_VAULT_PATH environment variable."
+
+    try:
+        note = obsidian_service.read_note(relative_path)
+        content_data = note["content"]
+
+        result = f"File: {note['path']}\n\n"
+        result += f"Frontmatter:\n{content_data['metadata']}\n\n"
+        result += f"Content:\n{content_data['content']}\n\n"
+        result += f"Tags: {', '.join(content_data['tags'])}\n"
+        result += f"Contains {len(content_data['wikilinks'])} wikilinks and {len(content_data['embeds'])} embeds"
+
+        return result
+
+    except FileNotFoundError:
+        return f"Note not found at '{relative_path}'. Please check the path is correct."
+    except Exception as e:
+        return f"Error reading note: {str(e)}"
+
+
+@tool
+async def obsidian_create_note(
+    title: str,
+    content: str,
+    folder: str = "notes",
+    tags: str = "",
+) -> str:
+    """Create a new note in your Obsidian vault.
+
+    Use this tool when you want to:
+    - Save research findings or ideas to your vault
+    - Create a new document with a specific title
+    - Write notes for future reference
+
+    Args:
+        title: The title of the note (will be used as filename)
+        content: The body content of the note
+        folder: The folder where to create the note (default: "notes")
+        tags: Comma-separated list of tags to add (default: "")
+
+    Returns:
+        Path to the created note
+    """
+    if not obsidian_enabled:
+        return "Obsidian integration is not configured. Please set OBSIDIAN_VAULT_PATH environment variable."
+
+    try:
+        # Parse tags from comma-separated string
+        tag_list = [tag.strip() for tag in tags.split(",") if tag.strip()]
+
+        relative_path = obsidian_service.create_note(
+            title=title,
+            content=content,
+            folder=folder,
+            tags=tag_list,
+        )
+
+        return f"Successfully created note: {relative_path}"
+
+    except Exception as e:
+        return f"Error creating note: {str(e)}"
+
+
+@tool
+def journal_get_today() -> str:
+    """Get today's daily journal note, including all tasks and log entries.
+
+    Use this tool when the user asks about:
+    - What's on their plate today
+    - Today's tasks or to-do list
+    - Today's journal entry
+    - What they've logged today
+
+    Returns:
+        The full content of today's daily note, or a message if it doesn't exist.
+    """
+    if not obsidian_enabled:
+        return "Obsidian integration is not configured."
+
+    try:
+        note = obsidian_service.get_daily_note()
+        if not note["found"]:
+            return f"No daily note found for {note['date']}. Use journal_add_task to create one."
+        return f"Daily note for {note['date']}:\n\n{note['content']}"
+    except Exception as e:
+        return f"Error reading daily note: {str(e)}"
+
+
+@tool
+def journal_get_tasks(date: str = "") -> str:
+    """Get tasks from a daily journal note.
+
+    Use this tool when the user asks about:
+    - Open or pending tasks for a day
+    - What tasks are done or not done
+    - Task status for today or a specific date
+
+    Args:
+        date: Date in YYYY-MM-DD format (optional, defaults to today)
+
+    Returns:
+        List of tasks with their completion status.
+    """
+    if not obsidian_enabled:
+        return "Obsidian integration is not configured."
+
+    try:
+        from datetime import datetime as dt
+
+        parsed_date = dt.strptime(date, "%Y-%m-%d") if date else None
+        result = obsidian_service.get_daily_tasks(parsed_date)
+
+        if not result["found"]:
+            return f"No daily note found for {result['date']}."
+
+        if not result["tasks"]:
+            return f"No tasks found in the {result['date']} note."
+
+        lines = [f"Tasks for {result['date']}:"]
+        for task in result["tasks"]:
+            status = "[x]" if task["done"] else "[ ]"
+            lines.append(f"- {status} {task['text']}")
+        return "\n".join(lines)
+    except Exception as e:
+        return f"Error reading tasks: {str(e)}"
+
+
+@tool
+def journal_add_task(task: str, date: str = "") -> str:
+    """Add a task to a daily journal note.
+
+    Use this tool when the user wants to:
+    - Add a task or to-do to today's note
+    - Remind themselves to do something
+    - Track a new item in their daily note
+
+    Args:
+        task: The task description to add
+        date: Date in YYYY-MM-DD format (optional, defaults to today)
+
+    Returns:
+        Confirmation of the added task.
+    """
+    if not obsidian_enabled:
+        return "Obsidian integration is not configured."
+
+    try:
+        from datetime import datetime as dt
+
+        parsed_date = dt.strptime(date, "%Y-%m-%d") if date else None
+        result = obsidian_service.add_task_to_daily_note(task, parsed_date)
+
+        if result["success"]:
+            note_date = date or dt.now().strftime("%Y-%m-%d")
+            extra = " (created new note)" if result["created_note"] else ""
+            return f"Added task '{task}' to {note_date}{extra}."
+        return "Failed to add task."
+    except Exception as e:
+        return f"Error adding task: {str(e)}"
+
+
+@tool
+def journal_complete_task(task: str, date: str = "") -> str:
+    """Mark a task as complete in a daily journal note.
+
+    Use this tool when the user wants to:
+    - Check off a task as done
+    - Mark something as completed
+    - Update task status in their daily note
+
+    Args:
+        task: The task text to mark complete (exact or partial match)
+        date: Date in YYYY-MM-DD format (optional, defaults to today)
+
+    Returns:
+        Confirmation that the task was marked complete.
+    """
+    if not obsidian_enabled:
+        return "Obsidian integration is not configured."
+
+    try:
+        from datetime import datetime as dt
+
+        parsed_date = dt.strptime(date, "%Y-%m-%d") if date else None
+        result = obsidian_service.complete_task_in_daily_note(task, parsed_date)
+
+        if result["success"]:
+            return f"Marked '{result['completed_task']}' as complete."
+        return f"Could not complete task: {result.get('error', 'unknown error')}"
+    except Exception as e:
+        return f"Error completing task: {str(e)}"
+
+
+@tool
+async def obsidian_create_task(
+    title: str,
+    content: str = "",
+    folder: str = "tasks",
+    due_date: str = "",
+    tags: str = "",
+) -> str:
+    """Create a new task note in your Obsidian vault.
+
+    Use this tool when you want to:
+    - Create a task to remember to do something
+    - Add a task with a due date
+    - Track tasks in your vault
+
+    Args:
+        title: The title of the task
+        content: The description of the task (optional)
+        folder: The folder to place the task (default: "tasks")
+        due_date: Due date in YYYY-MM-DD format (optional)
+        tags: Comma-separated list of tags to add (optional)
+
+    Returns:
+        Path to the created task note
+    """
+    if not obsidian_enabled:
+        return "Obsidian integration is not configured. Please set OBSIDIAN_VAULT_PATH environment variable."
+
+    try:
+        # Parse tags from comma-separated string
+        tag_list = [tag.strip() for tag in tags.split(",") if tag.strip()]
+
+        relative_path = obsidian_service.create_task(
+            title=title,
+            content=content,
+            folder=folder,
+            due_date=due_date or None,
+            tags=tag_list,
+        )
+
+        return f"Successfully created task: {relative_path}"
+
+    except Exception as e:
+        return f"Error creating task: {str(e)}"
+
+
+# Create tools list based on what's available
+tools = [get_current_date, simba_search, web_search]
+if ynab_enabled:
+    tools.extend(
+        [
+            ynab_budget_summary,
+            ynab_search_transactions,
+            ynab_category_spending,
+            ynab_insights,
+        ]
+    )
+if obsidian_enabled:
+    tools.extend(
+        [
+            obsidian_search_notes,
+            obsidian_read_note,
+            obsidian_create_note,
+            obsidian_create_task,
+            journal_get_today,
+            journal_get_tasks,
+            journal_add_task,
+            journal_complete_task,
+        ]
+    )
+
+# Llama 3.1 supports native function calling via OpenAI-compatible API
+main_agent = create_agent(model=model_with_fallback, tools=tools)

blueprints/rag/__init__.py

@@ -1,7 +1,7 @@
 from quart import Blueprint, jsonify
 from quart_jwt_extended import jwt_refresh_token_required
 
-from .logic import get_vector_store_stats, index_documents, vector_store
+from .logic import fetch_obsidian_documents, get_vector_store_stats, index_documents, index_obsidian_documents, vector_store
 from blueprints.users.decorators import admin_required
 
 rag_blueprint = Blueprint("rag_api", __name__, url_prefix="/api/rag")
@@ -45,3 +45,15 @@ async def trigger_reindex():
         return jsonify({"status": "success", "stats": stats})
     except Exception as e:
         return jsonify({"status": "error", "message": str(e)}), 500
+
+
+@rag_blueprint.post("/index-obsidian")
+@admin_required
+async def trigger_obsidian_index():
+    """Index all Obsidian markdown documents into vector store. Admin only."""
+    try:
+        result = await index_obsidian_documents()
+        stats = get_vector_store_stats()
+        return jsonify({"status": "success", "result": result, "stats": stats})
+    except Exception as e:
+        return jsonify({"status": "error", "message": str(e)}), 500

blueprints/rag/fetchers.py

@@ -1,8 +1,12 @@
 import os
 import tempfile
 
+from dotenv import load_dotenv
 import httpx
 
+# Load environment variables
+load_dotenv()
+
 
 class PaperlessNGXService:
     def __init__(self):

blueprints/rag/logic.py

@@ -1,12 +1,17 @@
 import datetime
 import os
 
+from dotenv import load_dotenv
 from langchain_chroma import Chroma
 from langchain_core.documents import Document
 from langchain_openai import OpenAIEmbeddings
 from langchain_text_splitters import RecursiveCharacterTextSplitter
 
 from .fetchers import PaperlessNGXService
+from utils.obsidian_service import ObsidianService
+
+# Load environment variables
+load_dotenv()
 
 embeddings = OpenAIEmbeddings(model="text-embedding-3-small")
 
@@ -54,12 +59,75 @@ async def fetch_documents_from_paperless_ngx() -> list[Document]:
 
 
 async def index_documents():
+    """Index Paperless-NGX documents into vector store."""
     documents = await fetch_documents_from_paperless_ngx()
 
     splits = text_splitter.split_documents(documents)
     await vector_store.aadd_documents(documents=splits)
 
 
+async def fetch_obsidian_documents() -> list[Document]:
+    """Fetch all markdown documents from Obsidian vault.
+
+    Returns:
+        List of LangChain Document objects with source='obsidian' metadata.
+    """
+    obsidian_service = ObsidianService()
+    documents = []
+
+    for md_path in obsidian_service.walk_vault():
+        try:
+            # Read markdown file
+            with open(md_path, "r", encoding="utf-8") as f:
+                content = f.read()
+
+            # Parse metadata
+            parsed = obsidian_service.parse_markdown(content, md_path)
+
+            # Create LangChain Document with obsidian source
+            document = Document(
+                page_content=parsed["content"],
+                metadata={
+                    "source": "obsidian",
+                    "filepath": parsed["filepath"],
+                    "tags": parsed["tags"],
+                    "created_at": parsed["metadata"].get("created_at"),
+                    **{k: v for k, v in parsed["metadata"].items() if k not in ["created_at", "created_by"]},
+                },
+            )
+            documents.append(document)
+
+        except Exception as e:
+            print(f"Error reading {md_path}: {e}")
+            continue
+
+    return documents
+
+
+async def index_obsidian_documents():
+    """Index all Obsidian markdown documents into vector store.
+
+    Deletes existing obsidian source chunks before re-indexing.
+    """
+    obsidian_service = ObsidianService()
+    documents = await fetch_obsidian_documents()
+
+    if not documents:
+        print("No Obsidian documents found to index")
+        return {"indexed": 0}
+
+    # Delete existing obsidian chunks
+    existing_results = vector_store.get(where={"source": "obsidian"})
+    if existing_results.get("ids"):
+        await vector_store.adelete(existing_results["ids"])
+
+    # Split and index documents
+    splits = text_splitter.split_documents(documents)
+    await vector_store.aadd_documents(documents=splits)
+
+    return {"indexed": len(documents)}
+
+
 async def query_vector_store(query: str):
     retrieved_docs = await vector_store.asimilarity_search(query, k=2)
     serialized = "\n\n".join(

blueprints/users/models.py

@@ -10,6 +10,7 @@ class User(Model):
     username = fields.CharField(max_length=255)
     password = fields.BinaryField(null=True)  # Hashed - nullable for OIDC users
     email = fields.CharField(max_length=100, unique=True)
+    whatsapp_number = fields.CharField(max_length=30, unique=True, null=True, index=True)
 
     # OIDC fields
     oidc_subject = fields.CharField(

blueprints/whatsapp/__init__.py

@@ -0,0 +1,254 @@
+import os
+import logging
+import asyncio
+import functools
+import time
+from collections import defaultdict
+from quart import Blueprint, request, jsonify, abort
+from twilio.request_validator import RequestValidator
+from twilio.twiml.messaging_response import MessagingResponse
+
+from blueprints.users.models import User
+from blueprints.conversation.logic import (
+    get_conversation_for_user,
+    add_message_to_conversation,
+    get_conversation_transcript,
+)
+from blueprints.conversation.agents import main_agent
+
+whatsapp_blueprint = Blueprint("whatsapp_api", __name__, url_prefix="/api/whatsapp")
+
+# Configure logging
+logger = logging.getLogger(__name__)
+
+# Rate limiting: per-number message timestamps
+# Format: {phone_number: [timestamp1, timestamp2, ...]}
+_rate_limit_store: dict[str, list[float]] = defaultdict(list)
+
+# Configurable via env: max messages per window (default: 10 per 60s)
+RATE_LIMIT_MAX = int(os.getenv("WHATSAPP_RATE_LIMIT_MAX", "10"))
+RATE_LIMIT_WINDOW = int(os.getenv("WHATSAPP_RATE_LIMIT_WINDOW", "60"))
+
+# Max message length to process (WhatsApp max is 4096, but we cap for LLM sanity)
+MAX_MESSAGE_LENGTH = 2000
+
+
+def _twiml_response(text: str) -> tuple[str, int]:
+    """Helper to return a TwiML MessagingResponse."""
+    resp = MessagingResponse()
+    resp.message(text)
+    return str(resp), 200
+
+
+def _check_rate_limit(phone_number: str) -> bool:
+    """Check if a phone number has exceeded the rate limit.
+
+    Returns True if the request is allowed, False if rate-limited.
+    Also cleans up expired entries.
+    """
+    now = time.monotonic()
+    cutoff = now - RATE_LIMIT_WINDOW
+
+    # Remove expired timestamps
+    timestamps = _rate_limit_store[phone_number]
+    _rate_limit_store[phone_number] = [t for t in timestamps if t > cutoff]
+
+    if len(_rate_limit_store[phone_number]) >= RATE_LIMIT_MAX:
+        return False
+
+    _rate_limit_store[phone_number].append(now)
+    return True
+
+
+def validate_twilio_request(f):
+    """Decorator to validate that the request comes from Twilio.
+
+    Validates the X-Twilio-Signature header using the TWILIO_AUTH_TOKEN.
+    Set TWILIO_WEBHOOK_URL if behind a reverse proxy (e.g., ngrok, Caddy)
+    so the validated URL matches what Twilio signed against.
+    Set TWILIO_SIGNATURE_VALIDATION=false to disable in development.
+    """
+    @functools.wraps(f)
+    async def decorated_function(*args, **kwargs):
+        if os.getenv("TWILIO_SIGNATURE_VALIDATION", "true").lower() == "false":
+            return await f(*args, **kwargs)
+
+        auth_token = os.getenv("TWILIO_AUTH_TOKEN")
+        if not auth_token:
+            logger.error("TWILIO_AUTH_TOKEN not set — rejecting request")
+            abort(403)
+
+        twilio_signature = request.headers.get("X-Twilio-Signature")
+        if not twilio_signature:
+            logger.warning("Missing X-Twilio-Signature header")
+            abort(403)
+
+        # Use configured webhook URL if behind a proxy, otherwise use request URL
+        url = os.getenv("TWILIO_WEBHOOK_URL") or request.url
+        form_data = await request.form
+
+        validator = RequestValidator(auth_token)
+        if not validator.validate(url, form_data, twilio_signature):
+            logger.warning(f"Invalid Twilio signature for URL: {url}")
+            abort(403)
+
+        return await f(*args, **kwargs)
+    return decorated_function
+
+
+@whatsapp_blueprint.route("/webhook", methods=["POST"])
+@validate_twilio_request
+async def webhook():
+    """
+    Handle incoming WhatsApp messages from Twilio.
+    """
+    form_data = await request.form
+    from_number = form_data.get("From") # e.g., "whatsapp:+1234567890"
+    body = form_data.get("Body")
+
+    if not from_number or not body:
+        return _twiml_response("Invalid message received.") if from_number else ("Missing From or Body", 400)
+
+    # Strip whitespace and check for empty body
+    body = body.strip()
+    if not body:
+        return _twiml_response("I received an empty message. Please send some text!")
+
+    # Rate limiting
+    if not _check_rate_limit(from_number):
+        logger.warning(f"Rate limit exceeded for {from_number}")
+        return _twiml_response("You're sending messages too quickly. Please wait a moment and try again.")
+
+    # Truncate overly long messages
+    if len(body) > MAX_MESSAGE_LENGTH:
+        body = body[:MAX_MESSAGE_LENGTH]
+        logger.info(f"Truncated long message from {from_number} to {MAX_MESSAGE_LENGTH} chars")
+
+    logger.info(f"Received WhatsApp message from {from_number}: {body[:100]}")
+
+    # Identify or create user
+    user = await User.filter(whatsapp_number=from_number).first()
+
+    if not user:
+        # Check if number is in allowlist
+        allowed_numbers = os.getenv("ALLOWED_WHATSAPP_NUMBERS", "").split(",")
+        if from_number not in allowed_numbers and "*" not in allowed_numbers:
+            return _twiml_response("Sorry, you are not authorized to use this service.")
+
+        # Create a new user for this WhatsApp number
+        username = f"wa_{from_number.split(':')[-1]}"
+        try:
+            user = await User.create(
+                username=username,
+                email=f"{username}@whatsapp.simbarag.local",
+                whatsapp_number=from_number,
+                auth_provider="whatsapp"
+            )
+            logger.info(f"Created new user for WhatsApp: {username}")
+        except Exception as e:
+            logger.error(f"Failed to create user for {from_number}: {e}")
+            return _twiml_response("Sorry, something went wrong setting up your account. Please try again later.")
+
+    # Get or create a conversation for this user
+    try:
+        conversation = await get_conversation_for_user(user=user)
+        await conversation.fetch_related("messages")
+    except Exception as e:
+        logger.error(f"Failed to get conversation for user {user.username}: {e}")
+        return _twiml_response("Sorry, something went wrong. Please try again later.")
+
+    # Add user message to conversation
+    await add_message_to_conversation(
+        conversation=conversation,
+        message=body,
+        speaker="user",
+        user=user,
+    )
+
+    # Get transcript for context
+    transcript = await get_conversation_transcript(user=user, conversation=conversation)
+
+    # Build messages payload for LangChain agent with system prompt and conversation history
+    try:
+        # System prompt with Simba's facts and medical information
+        system_prompt = """You are a helpful cat assistant named Simba that understands veterinary terms. When there are questions to you specifically, they are referring to Simba the cat. Answer the user in as if you were a cat named Simba. Don't act too catlike. Be assertive.
+
+SIMBA FACTS (as of January 2026):
+- Name: Simba
+- Species: Feline (Domestic Short Hair / American Short Hair)
+- Sex: Male, Neutered
+- Date of Birth: August 8, 2016 (approximately 9 years 5 months old)
+- Color: Orange
+- Current Weight: 16 lbs (as of 1/8/2026)
+- Owner: Ryan Chen
+- Location: Long Island City, NY
+- Veterinarian: Court Square Animal Hospital
+
+Medical Conditions:
+- Hypertrophic Cardiomyopathy (HCM): Diagnosed 12/11/2025. Concentric left ventricular hypertrophy with no left atrial dilation. Grade II-III/VI systolic heart murmur. No cardiac medications currently needed. Must avoid Domitor, acepromazine, and ketamine during anesthesia.
+- Dental Issues: Prior extraction of teeth 307 and 407 due to resorption. Tooth 107 extracted on 1/8/2026. Early resorption lesions present on teeth 207, 309, and 409.
+
+Recent Medical Events:
+- 1/8/2026: Dental cleaning and tooth 107 extraction. Prescribed Onsior for 3 days. Oravet sealant applied.
+- 12/11/2025: Echocardiogram confirming HCM diagnosis. Pre-op bloodwork was normal.
+- 12/1/2025: Visited for decreased appetite/nausea. Received subcutaneous fluids and Cerenia.
+
+Diet & Lifestyle:
+- Diet: Hill's I/D wet and dry food
+- Supplements: Plaque Off
+- Indoor only cat, only pet in the household
+
+Upcoming Appointments:
+- Rabies Vaccine: Due 2/19/2026
+- Routine Examination: Due 6/1/2026
+- FVRCP-3yr Vaccine: Due 10/2/2026
+
+IMPORTANT: When users ask factual questions about Simba's health, medical history, veterinary visits, medications, weight, or any information that would be in documents, you MUST use the simba_search tool to retrieve accurate information before answering. Do not rely on general knowledge - always search the documents for factual questions.
+
+BUDGET & FINANCE (YNAB Integration):
+You have access to Ryan's budget data through YNAB (You Need A Budget). When users ask about financial matters, use the appropriate YNAB tools:
+- Use ynab_budget_summary for overall budget health and status questions
+- Use ynab_search_transactions to find specific purchases or spending at particular stores
+- Use ynab_category_spending to analyze spending by category for a month
+- Use ynab_insights to provide spending trends, patterns, and recommendations
+Always use these tools when asked about budgets, spending, transactions, or financial health."""
+
+        # Get last 10 messages for conversation history
+        messages = await conversation.messages.all()
+        recent_messages = list(messages)[-10:]
+
+        # Build messages payload
+        messages_payload = [{"role": "system", "content": system_prompt}]
+
+        # Add recent conversation history (exclude the message we just added)
+        for msg in recent_messages[:-1]:
+            role = "user" if msg.speaker == "user" else "assistant"
+            messages_payload.append({"role": role, "content": msg.text})
+
+        # Add current query
+        messages_payload.append({"role": "user", "content": body})
+
+        # Invoke LangChain agent
+        logger.info(f"Invoking LangChain agent with {len(messages_payload)} messages")
+        response = await main_agent.ainvoke({"messages": messages_payload})
+        response_text = response.get("messages", [])[-1].content
+
+        # Log YNAB availability
+        if os.getenv("YNAB_ACCESS_TOKEN"):
+            logger.info("YNAB integration is available for this conversation")
+        else:
+            logger.info("YNAB integration is not configured")
+
+    except Exception as e:
+        logger.error(f"Error invoking agent: {e}")
+        response_text = "Sorry, I'm having trouble thinking right now. 😿"
+
+    # Add Simba's response to conversation
+    await add_message_to_conversation(
+        conversation=conversation,
+        message=response_text,
+        speaker="simba",
+        user=user,
+    )
+
+    return _twiml_response(response_text)

config/oidc_config.py

@@ -7,6 +7,10 @@ from typing import Dict, Any
 from authlib.jose import jwt
 from authlib.jose.errors import JoseError
 import httpx
+from dotenv import load_dotenv
+
+# Load environment variables
+load_dotenv()
 
 
 class OIDCConfig:

docker-compose.dev.yml

@@ -1,71 +0,0 @@
-services:
-  postgres:
-    image: postgres:16-alpine
-    environment:
-      - POSTGRES_USER=raggr
-      - POSTGRES_PASSWORD=raggr_dev_password
-      - POSTGRES_DB=raggr
-    ports:
-      - "5432:5432"
-    volumes:
-      - postgres_data:/var/lib/postgresql/data
-    healthcheck:
-      test: ["CMD-SHELL", "pg_isready -U raggr"]
-      interval: 5s
-      timeout: 5s
-      retries: 5
-
-  # raggr service disabled - run locally for development
-  # raggr:
-  #   build:
-  #     context: .
-  #     dockerfile: Dockerfile.dev
-  #   image: torrtle/simbarag:dev
-  #   ports:
-  #     - "8080:8080"
-  #   env_file:
-  #     - .env
-  #   environment:
-  #     - PAPERLESS_TOKEN=${PAPERLESS_TOKEN}
-  #     - BASE_URL=${BASE_URL}
-  #     - OLLAMA_URL=${OLLAMA_URL:-http://localhost:11434}
-  #     - CHROMADB_PATH=/app/data/chromadb
-  #     - OPENAI_API_KEY=${OPENAI_API_KEY}
-  #     - JWT_SECRET_KEY=${JWT_SECRET_KEY}
-  #     - OIDC_ISSUER=${OIDC_ISSUER}
-  #     - OIDC_CLIENT_ID=${OIDC_CLIENT_ID}
-  #     - OIDC_CLIENT_SECRET=${OIDC_CLIENT_SECRET}
-  #     - OIDC_REDIRECT_URI=${OIDC_REDIRECT_URI}
-  #     - OIDC_USE_DISCOVERY=${OIDC_USE_DISCOVERY:-true}
-  #     - DATABASE_URL=postgres://raggr:raggr_dev_password@postgres:5432/raggr
-  #     - FLASK_ENV=development
-  #     - PYTHONUNBUFFERED=1
-  #     - NODE_ENV=development
-  #     - TAVILY_KEY=${TAVILIY_KEY}
-  #   depends_on:
-  #     postgres:
-  #       condition: service_healthy
-  #   volumes:
-  #     - chromadb_data:/app/data/chromadb
-  #     - ./migrations:/app/migrations  # Bind mount for migrations (bidirectional)
-  #   develop:
-  #     watch:
-  #       # Sync+restart on any file change in root directory
-  #       - action: sync+restart
-  #         path: .
-  #         target: /app
-  #         ignore:
-  #           - __pycache__/
-  #           - "*.pyc"
-  #           - "*.pyo"
-  #           - "*.pyd"
-  #           - .git/
-  #           - chromadb/
-  #           - node_modules/
-  #           - raggr-frontend/dist/
-  #           - docs/
-  #           - .venv/
-
-volumes:
-  chromadb_data:
-  postgres_data:

docker-compose.yml

@@ -32,18 +32,35 @@ services:
       - CHROMADB_PATH=/app/data/chromadb
       - OPENAI_API_KEY=${OPENAI_API_KEY}
       - JWT_SECRET_KEY=${JWT_SECRET_KEY}
+      - LLAMA_SERVER_URL=${LLAMA_SERVER_URL}
+      - LLAMA_MODEL_NAME=${LLAMA_MODEL_NAME}
       - OIDC_ISSUER=${OIDC_ISSUER}
       - OIDC_CLIENT_ID=${OIDC_CLIENT_ID}
       - OIDC_CLIENT_SECRET=${OIDC_CLIENT_SECRET}
       - OIDC_REDIRECT_URI=${OIDC_REDIRECT_URI}
       - OIDC_USE_DISCOVERY=${OIDC_USE_DISCOVERY:-true}
       - DATABASE_URL=${DATABASE_URL:-postgres://raggr:changeme@postgres:5432/raggr}
-      - TAVILY_KEY=${TAVILIY_KEY}
+      - TAVILY_API_KEY=${TAVILIY_API_KEY}
+      - YNAB_ACCESS_TOKEN=${YNAB_ACCESS_TOKEN}
+      - YNAB_BUDGET_ID=${YNAB_BUDGET_ID}
+      - TWILIO_ACCOUNT_SID=${TWILIO_ACCOUNT_SID}
+      - TWILIO_AUTH_TOKEN=${TWILIO_AUTH_TOKEN}
+      - TWILIO_WHATSAPP_NUMBER=${TWILIO_WHATSAPP_NUMBER}
+      - ALLOWED_WHATSAPP_NUMBERS=${ALLOWED_WHATSAPP_NUMBERS}
+      - TWILIO_SIGNATURE_VALIDATION=${TWILIO_SIGNATURE_VALIDATION:-true}
+      - TWILIO_WEBHOOK_URL=${TWILIO_WEBHOOK_URL:-}
+      - OBSIDIAN_AUTH_TOKEN=${OBSIDIAN_AUTH_TOKEN}
+      - OBSIDIAN_VAULT_ID=${OBSIDIAN_VAULT_ID}
+      - OBSIDIAN_E2E_PASSWORD=${OBSIDIAN_E2E_PASSWORD}
+      - OBSIDIAN_DEVICE_NAME=${OBSIDIAN_DEVICE_NAME}
+      - OBSIDIAN_CONTINUOUS_SYNC=${OBSIDIAN_CONTINUOUS_SYNC:-false}
+      - OBSIDIAN_VAULT_PATH=${OBSIDIAN_VAULT_PATH:-/app/data/obsidian}
     depends_on:
       postgres:
         condition: service_healthy
     volumes:
       - chromadb_data:/app/data/chromadb
+      - ./obvault:/app/data/obsidian
     restart: unless-stopped
 
 volumes:

main.py

@@ -225,6 +225,10 @@ def filter_indexed_files(docs):
 def reindex():
     with sqlite3.connect("database/visited.db") as conn:
         c = conn.cursor()
+        # Ensure the table exists before trying to delete from it
+        c.execute(
+            "CREATE TABLE IF NOT EXISTS indexed_documents (id INTEGER PRIMARY KEY AUTOINCREMENT, paperless_id INTEGER)"
+        )
         c.execute("DELETE FROM indexed_documents")
         conn.commit()
 

migrations/models/2_20260228125713_add_whatsapp_number.py

@@ -0,0 +1,42 @@
+from tortoise import BaseDBAsyncClient
+
+RUN_IN_TRANSACTION = True
+
+
+async def upgrade(db: BaseDBAsyncClient) -> str:
+    return """
+        ALTER TABLE "users" ADD "whatsapp_number" VARCHAR(20) UNIQUE;"""
+
+
+async def downgrade(db: BaseDBAsyncClient) -> str:
+    return """
+        DROP INDEX IF EXISTS "uid_users_whatsap_e6b586";
+        ALTER TABLE "users" DROP COLUMN "whatsapp_number";"""
+
+
+MODELS_STATE = (
+    "eJztmm1v4jgQx78Kyquu1KtatnRX1emkQOkttwuceNinXhWZxECuiZ2NnaWo6nc/2yTESR"
+    "wgFCjs8aYtYw+2fx5n/p70SXOxBR1yVsPoJ/QJoDZG2nXpSUPAhewPZftpSQOeF7dyAwUD"
+    "RziYUk/RAgaE+sCkrHEIHAKZyYLE9G0vHAwFjsON2GQdbTSKTQGyfwTQoHgE6Rj6rOHunp"
+    "ltZMFHSKKP3oMxtKFjJeZtW3xsYTfo1BO2fr9xcyt68uEGhomdwEVxb29KxxjNuweBbZ1x"
+    "H942ggj6gEJLWgafZbjsyDSbMTNQP4DzqVqxwYJDEDgchvb7MEAmZ1ASI/Efl39oBfAw1B"
+    "ytjShn8fQ8W1W8ZmHV+FC1D3rn5O3VG7FKTOjIF42CiPYsHAEFM1fBNQYpfmdQ1sbAV6OM"
+    "+qdgsomugzEyxBzjGIpARoDWo6a54NFwIBrRMftYrlQWYPysdwRJ1kugxCyuZ1HfCpvKsz"
+    "aONEZo+pAv2QA0C/KGtVDbhWqYSc8UUit0PYv+2FPAbA1WGznT8BAs4NtrNOvdnt78m6/E"
+    "JeSHIxDpvTpvKQvrNGU9uUptxfxLSl8avQ8l/rH0vd2qp2N/3q/3XeNzAgHFBsITA1jSeY"
+    "2sEZjExgaetebGJj2PG/uqGxtOXtpXAn2jWAaRXF6QRsK57XAT108aPPUOH5Q5g8PIwrvF"
+    "PrRH6COcCoQNNg+ATFWyCEVHP/yafYUWW+NZ+GAyVyNyULDVsTVBOsueerem39Q1wXAAzI"
+    "cJ8C0jB6YLCQEjSLJAq6Hn7ccOdObSTM1SFnDN2Tfu51Mlj61ghctYYpSgl21yy27aAhBb"
+    "txWOzUdaQGeJCpYgriaGDXkjj6L4oEUxhY+KlN9jVjXKqP+hiOJFqbz+tZfI4pH0PWnqX9"
+    "8kMvmnduvPqLsklWuf2tWjQv4VhVRWIRMPggeVGOAXyDoK3IwUSOyu5P7KR0frd+ud6xLP"
+    "6P+gbqNZ1a9LxHYHQFttixO3zIvzFS6ZF+e5d0zelDpAcqIp9phXuG7ymX+gEtZMFbxeKG"
+    "XT9bO9pbhU0yrCpai23aaSE3cGhXSL7hL5Wo0f7aM2O3xtxvexaNFS9jkUjbaDwqUHCJlg"
+    "XxGZVRsBf6qGKXulYA6mdHb/2dcrvQpeletVWW4xZNVGS+98U0veqL8ct9VvvbqeogtdYD"
+    "tFonTusJkQXX7iNxmgF+eriZ5FqicjeyZjQAl7pBtMSQ7yZKYapsJ1LazpUN0t1fIqUMv5"
+    "TMsZpNi2TIMEg3+hqbiM5fNM+x0izG08Q9n1aGx4Pv5pW8UCNOO4u8SkOdgEzgsye5JrZZ"
+    "UgreQHaSUTpI4FPGPk48BTlEX/6rZbaqQptxTQPmKrvLNsk56WHJvQ+63hvbvfjmriK19c"
+    "m0mXYVJpin/BsTbzP6nNHN9e/hIbO385krljL3uzlPlXnc28Xtpnfb/b10o69G1zrCnKEW"
+    "HL6aKCBIj77E1FooFy3nAoCxIccyoYwp1/1XuJeLn3W/ni8t3l+7dXl+9ZFzGTueXdgodB"
+    "o9VbUoDgB0FZNczXepLLsfwQS2d2NIoI5ln3wwS4lesxG5FCpEjv+RJZcnkteby1Qs7G5H"
+    "GBbLv59PL8Hy/ZG1k="
+)

pyproject.toml

@@ -20,7 +20,7 @@ dependencies = [
     "pony>=0.7.19",
     "flask-login>=0.6.3",
     "quart>=0.20.0",
-    "tortoise-orm>=0.25.1",
+    "tortoise-orm>=0.25.1,<1.0.0",
     "quart-jwt-extended>=0.1.0",
     "pre-commit>=4.3.0",
     "tortoise-orm-stubs>=1.0.2",
@@ -34,6 +34,9 @@ dependencies = [
     "langchain-community>=0.4.1",
     "jq>=1.10.0",
     "tavily-python>=0.7.17",
+    "ynab>=1.3.0",
+    "ollama>=0.6.1",
+    "twilio>=9.10.2",
 ]
 
 [tool.aerich]

raggr-frontend/src/App.css

@@ -1,7 +1,170 @@
 @import "tailwindcss";
+@import url('https://fonts.googleapis.com/css2?family=Nunito:wght@400;500;600;700&family=Playfair+Display:ital,wght@0,600;0,700;1,600&display=swap');
+
+@theme {
+  --color-cream: #FBF7F0;
+  --color-cream-dark: #F3EDE2;
+  --color-warm-white: #FFFDF9;
+  --color-amber-glow: #E8943A;
+  --color-amber-soft: #F5C882;
+  --color-amber-pale: #FFF0D6;
+  --color-forest: #2D5A3D;
+  --color-forest-light: #3D763A;
+  --color-forest-pale: #E8F5E4;
+  --color-charcoal: #2C2420;
+  --color-warm-gray: #8A7E74;
+  --color-sand: #D4C5B0;
+  --color-sand-light: #E8DED0;
+  --color-blush: #F2D1B3;
+  --color-sidebar-bg: #2C2420;
+  --color-sidebar-hover: #3D352F;
+  --color-sidebar-active: #4A3F38;
+  --font-display: 'Playfair Display', Georgia, serif;
+  --font-body: 'Nunito', -apple-system, BlinkMacSystemFont, sans-serif;
+}
+
+* {
+  box-sizing: border-box;
+}
 
 body {
   margin: 0;
-	font-family: Inter, Avenir, Helvetica, Arial, sans-serif;
+  font-family: var(--font-body);
-	background-color: #F9F5EB;
+  background-color: var(--color-cream);
+  color: var(--color-charcoal);
+  -webkit-font-smoothing: antialiased;
+  -moz-osx-font-smoothing: grayscale;
+}
+
+/* Scrollbar styling */
+::-webkit-scrollbar {
+  width: 6px;
+}
+::-webkit-scrollbar-track {
+  background: transparent;
+}
+::-webkit-scrollbar-thumb {
+  background: var(--color-sand);
+  border-radius: 3px;
+}
+::-webkit-scrollbar-thumb:hover {
+  background: var(--color-warm-gray);
+}
+
+/* Markdown content styling in answer bubbles */
+.markdown-content h1,
+.markdown-content h2,
+.markdown-content h3 {
+  font-family: var(--font-display);
+  font-weight: 600;
+  margin-top: 1em;
+  margin-bottom: 0.5em;
+  line-height: 1.3;
+}
+.markdown-content h1 { font-size: 1.25rem; }
+.markdown-content h2 { font-size: 1.1rem; }
+.markdown-content h3 { font-size: 1rem; }
+
+.markdown-content p {
+  margin: 0.5em 0;
+  line-height: 1.65;
+}
+
+.markdown-content ul,
+.markdown-content ol {
+  padding-left: 1.5em;
+  margin: 0.5em 0;
+}
+
+.markdown-content li {
+  margin: 0.25em 0;
+  line-height: 1.6;
+}
+
+.markdown-content code {
+  background: rgba(0, 0, 0, 0.06);
+  padding: 0.15em 0.4em;
+  border-radius: 4px;
+  font-size: 0.88em;
+  font-family: 'SF Mono', 'Fira Code', monospace;
+}
+
+.markdown-content pre {
+  background: var(--color-charcoal);
+  color: #F3EDE2;
+  padding: 1em;
+  border-radius: 8px;
+  overflow-x: auto;
+  margin: 0.75em 0;
+}
+
+.markdown-content pre code {
+  background: none;
+  padding: 0;
+  color: inherit;
+}
+
+.markdown-content a {
+  color: var(--color-forest);
+  text-decoration: underline;
+  text-underline-offset: 2px;
+}
+
+.markdown-content blockquote {
+  border-left: 3px solid var(--color-amber-glow);
+  padding-left: 1em;
+  margin: 0.75em 0;
+  color: var(--color-warm-gray);
+  font-style: italic;
+}
+
+/* Loading skeleton animation */
+@keyframes shimmer {
+  0% { background-position: -200% 0; }
+  100% { background-position: 200% 0; }
+}
+
+.skeleton-shimmer {
+  background: linear-gradient(
+    90deg,
+    var(--color-sand-light) 25%,
+    var(--color-cream) 50%,
+    var(--color-sand-light) 75%
+  );
+  background-size: 200% 100%;
+  animation: shimmer 1.8s ease-in-out infinite;
+}
+
+/* Fade-in animation for messages */
+@keyframes fadeSlideUp {
+  from {
+    opacity: 0;
+    transform: translateY(12px);
+  }
+  to {
+    opacity: 1;
+    transform: translateY(0);
+  }
+}
+
+.message-enter {
+  animation: fadeSlideUp 0.35s ease-out forwards;
+}
+
+/* Subtle pulse for loading dots */
+@keyframes catPulse {
+  0%, 80%, 100% { opacity: 0.3; transform: scale(0.8); }
+  40% { opacity: 1; transform: scale(1); }
+}
+
+.loading-dot {
+  animation: catPulse 1.4s ease-in-out infinite;
+}
+.loading-dot:nth-child(2) { animation-delay: 0.2s; }
+.loading-dot:nth-child(3) { animation-delay: 0.4s; }
+
+/* Textarea focus glow */
+textarea:focus {
+  outline: none;
+  box-shadow: 0 0 0 2px var(--color-amber-soft);
 }

raggr-frontend/src/App.tsx

@@ -5,6 +5,7 @@ import { AuthProvider } from "./contexts/AuthContext";
 import { ChatScreen } from "./components/ChatScreen";
 import { LoginScreen } from "./components/LoginScreen";
 import { conversationService } from "./api/conversationService";
+import catIcon from "./assets/cat.png";
 
 const AppContainer = () => {
   const [isAuthenticated, setAuthenticated] = useState<boolean>(false);
@@ -44,8 +45,15 @@ const AppContainer = () => {
   // Show loading state while checking authentication
   if (isChecking) {
     return (
-      <div className="h-screen flex items-center justify-center bg-white/85">
+      <div className="h-screen flex flex-col items-center justify-center bg-cream gap-4">
-        <div className="text-xl">Loading...</div>
+        <img
+          src={catIcon}
+          alt="Simba"
+          className="w-16 h-16 animate-bounce"
+        />
+        <p className="text-warm-gray font-medium text-lg tracking-wide">
+          waking up simba...
+        </p>
       </div>
     );
   }

raggr-frontend/src/api/conversationService.ts

@@ -1,5 +1,13 @@
 import { userService } from "./userService";
 
+export type SSEEvent =
+  | { type: "tool_start"; tool: string }
+  | { type: "tool_end"; tool: string }
+  | { type: "response"; message: string }
+  | { type: "error"; message: string };
+
+export type SSEEventCallback = (event: SSEEvent) => void;
+
 interface Message {
   id: string;
   text: string;
@@ -35,12 +43,14 @@ class ConversationService {
   async sendQuery(
     query: string,
     conversation_id: string,
+    signal?: AbortSignal,
   ): Promise<QueryResponse> {
     const response = await userService.fetchWithRefreshToken(
       `${this.conversationBaseUrl}/query`,
       {
         method: "POST",
         body: JSON.stringify({ query, conversation_id }),
+        signal,
       },
     );
 
@@ -110,6 +120,59 @@ class ConversationService {
 
     return await response.json();
   }
+
+  async streamQuery(
+    query: string,
+    conversation_id: string,
+    onEvent: SSEEventCallback,
+    signal?: AbortSignal,
+  ): Promise<void> {
+    const response = await userService.fetchWithRefreshToken(
+      `${this.conversationBaseUrl}/stream-query`,
+      {
+        method: "POST",
+        body: JSON.stringify({ query, conversation_id }),
+        signal,
+      },
+    );
+
+    if (!response.ok) {
+      throw new Error("Failed to stream query");
+    }
+
+    await this._readSSEStream(response, onEvent);
+  }
+
+  private async _readSSEStream(
+    response: Response,
+    onEvent: SSEEventCallback,
+  ): Promise<void> {
+    const reader = response.body!.getReader();
+    const decoder = new TextDecoder();
+    let buffer = "";
+
+    while (true) {
+      const { done, value } = await reader.read();
+      if (done) break;
+
+      buffer += decoder.decode(value, { stream: true });
+      const parts = buffer.split("\n\n");
+      buffer = parts.pop() ?? "";
+
+      for (const part of parts) {
+        const line = part.trim();
+        if (!line.startsWith("data: ")) continue;
+        const data = line.slice(6);
+        if (data === "[DONE]") return;
+        try {
+          const event = JSON.parse(data) as SSEEvent;
+          onEvent(event);
+        } catch {
+          // ignore malformed events
+        }
+      }
+    }
+  }
 }
 
 export const conversationService = new ConversationService();

raggr-frontend/src/components/ChatScreen.tsx

@@ -2,13 +2,14 @@ import { useEffect, useState, useRef } from "react";
 import { conversationService } from "../api/conversationService";
 import { QuestionBubble } from "./QuestionBubble";
 import { AnswerBubble } from "./AnswerBubble";
+import { ToolBubble } from "./ToolBubble";
 import { MessageInput } from "./MessageInput";
 import { ConversationList } from "./ConversationList";
 import catIcon from "../assets/cat.png";
 
 type Message = {
   text: string;
-  speaker: "simba" | "user";
+  speaker: "simba" | "user" | "tool";
 };
 
 type QuestionAnswer = {
@@ -25,6 +26,24 @@ type ChatScreenProps = {
   setAuthenticated: (isAuth: boolean) => void;
 };
 
+const TOOL_MESSAGES: Record<string, string> = {
+  simba_search: "🔍 Searching Simba's records...",
+  web_search: "🌐 Searching the web...",
+  get_current_date: "📅 Checking today's date...",
+  ynab_budget_summary: "💰 Checking budget summary...",
+  ynab_search_transactions: "💳 Looking up transactions...",
+  ynab_category_spending: "📊 Analyzing category spending...",
+  ynab_insights: "📈 Generating budget insights...",
+  obsidian_search_notes: "📝 Searching notes...",
+  obsidian_read_note: "📖 Reading note...",
+  obsidian_create_note: "✏️ Saving note...",
+  obsidian_create_task: "✅ Creating task...",
+  journal_get_today: "📔 Reading today's journal...",
+  journal_get_tasks: "📋 Getting tasks...",
+  journal_add_task: "➕ Adding task...",
+  journal_complete_task: "✔️ Completing task...",
+};
+
 export const ChatScreen = ({ setAuthenticated }: ChatScreenProps) => {
   const [query, setQuery] = useState<string>("");
   const [answer, setAnswer] = useState<string>("");
@@ -43,12 +62,26 @@ export const ChatScreen = ({ setAuthenticated }: ChatScreenProps) => {
   const [isLoading, setIsLoading] = useState<boolean>(false);
 
   const messagesEndRef = useRef<HTMLDivElement>(null);
+  const isMountedRef = useRef<boolean>(true);
+  const abortControllerRef = useRef<AbortController | null>(null);
   const simbaAnswers = ["meow.", "hiss...", "purrrrrr", "yowOWROWWowowr"];
 
   const scrollToBottom = () => {
     messagesEndRef.current?.scrollIntoView({ behavior: "smooth" });
   };
 
+  // Cleanup effect to handle component unmounting
+  useEffect(() => {
+    isMountedRef.current = true;
+    return () => {
+      isMountedRef.current = false;
+      // Abort any pending requests when component unmounts
+      if (abortControllerRef.current) {
+        abortControllerRef.current.abort();
+      }
+    };
+  }, []);
+
   const handleSelectConversation = (conversation: Conversation) => {
     setShowConversations(false);
     setSelectedConversation(conversation);
@@ -80,8 +113,6 @@ export const ChatScreen = ({ setAuthenticated }: ChatScreenProps) => {
       }));
       setConversations(parsedConversations);
       setSelectedConversation(parsedConversations[0]);
-      console.log(parsedConversations);
-      console.log("JELLYFISH@");
     } catch (error) {
       console.error("Failed to load messages:", error);
     }
@@ -106,8 +137,6 @@ export const ChatScreen = ({ setAuthenticated }: ChatScreenProps) => {
 
   useEffect(() => {
     const loadMessages = async () => {
-      console.log(selectedConversation);
-      console.log("JELLYFISH");
       if (selectedConversation == null) return;
       try {
         const conversation = await conversationService.getConversation(
@@ -140,7 +169,6 @@ export const ChatScreen = ({ setAuthenticated }: ChatScreenProps) => {
     setIsLoading(true);
 
     if (simbaMode) {
-      console.log("simba mode activated");
       const randomIndex = Math.floor(Math.random() * simbaAnswers.length);
       const randomElement = simbaAnswers[randomIndex];
       setAnswer(randomElement);
@@ -156,26 +184,49 @@ export const ChatScreen = ({ setAuthenticated }: ChatScreenProps) => {
       return;
     }
 
+    // Create a new AbortController for this request
+    const abortController = new AbortController();
+    abortControllerRef.current = abortController;
+
     try {
-      const result = await conversationService.sendQuery(
+      await conversationService.streamQuery(
         query,
         selectedConversation.id,
+        (event) => {
+          if (!isMountedRef.current) return;
+          if (event.type === "tool_start") {
+            const friendly =
+              TOOL_MESSAGES[event.tool] ?? `🔧 Using ${event.tool}...`;
+            setMessages((prev) => prev.concat([{ text: friendly, speaker: "tool" }]));
+          } else if (event.type === "response") {
+            setMessages((prev) =>
+              prev.concat([{ text: event.message, speaker: "simba" }]),
             );
-      setQuestionsAnswers(
+          } else if (event.type === "error") {
-        questionsAnswers.concat([{ question: query, answer: result.response }]),
+            console.error("Stream error:", event.message);
-      );
+          }
-      setMessages(
+        },
-        currMessages.concat([{ text: result.response, speaker: "simba" }]),
+        abortController.signal,
       );
     } catch (error) {
+      // Ignore abort errors (these are intentional cancellations)
+      if (error instanceof Error && error.name === "AbortError") {
+        console.log("Request was aborted");
+      } else {
         console.error("Failed to send query:", error);
         // If session expired, redirect to login
         if (error instanceof Error && error.message.includes("Session expired")) {
           setAuthenticated(false);
         }
+      }
     } finally {
+      // Only update loading state if component is still mounted
+      if (isMountedRef.current) {
         setIsLoading(false);
       }
+      // Clear the abort controller reference
+      abortControllerRef.current = null;
+    }
   };
 
   const handleQueryChange = (event: React.ChangeEvent<HTMLTextAreaElement>) => {
@@ -190,43 +241,62 @@ export const ChatScreen = ({ setAuthenticated }: ChatScreenProps) => {
     }
   };
 
+  const handleLogout = () => {
+    localStorage.removeItem("access_token");
+    localStorage.removeItem("refresh_token");
+    setAuthenticated(false);
+  };
+
   return (
-    <div className="h-screen flex flex-row bg-[#F9F5EB]">
+    <div className="h-screen flex flex-row bg-cream">
-      {/* Sidebar - Expanded */}
+      {/* Sidebar */}
       <aside
-        className={`hidden md:flex md:flex-col bg-[#F9F5EB] border-r border-gray-200 p-4 overflow-y-auto transition-all duration-300 ${sidebarCollapsed ? "w-20" : "w-64"}`}
+        className={`hidden md:flex md:flex-col bg-sidebar-bg transition-all duration-300 ease-in-out ${
+          sidebarCollapsed ? "w-[68px]" : "w-72"
+        }`}
       >
         {!sidebarCollapsed ? (
-          <div className="bg-[#F9F5EB]">
+          <div className="flex flex-col h-full">
-            <div className="flex flex-row items-center gap-2 mb-6">
+            {/* Sidebar header */}
+            <div className="flex items-center gap-3 px-5 py-5 border-b border-white/10">
               <img
                 src={catIcon}
                 alt="Simba"
-                className="cursor-pointer hover:opacity-80"
+                className="w-9 h-9 cursor-pointer hover:scale-110 transition-transform duration-200 flex-shrink-0"
                 onClick={() => setSidebarCollapsed(true)}
               />
-              <h2 className="text-3xl bg-[#F9F5EB] font-semibold">asksimba!</h2>
+              <h2 className="font-[family-name:var(--font-display)] text-xl font-bold text-cream tracking-tight">
+                asksimba
+              </h2>
             </div>
+
+            {/* Conversations */}
+            <div className="flex-1 overflow-y-auto px-3 py-3">
               <ConversationList
                 conversations={conversations}
                 onCreateNewConversation={handleCreateNewConversation}
                 onSelectConversation={handleSelectConversation}
+                selectedId={selectedConversation?.id}
               />
-            <div className="mt-auto pt-4">
+            </div>
+
+            {/* Logout */}
+            <div className="px-3 pb-4 pt-2 border-t border-white/10">
               <button
-                className="w-full p-2 border border-red-400 bg-red-200 hover:bg-red-400 cursor-pointer rounded-md text-sm"
+                className="w-full py-2.5 px-3 text-sm text-cream/60 hover:text-cream hover:bg-white/5
-                onClick={() => setAuthenticated(false)}
+                  rounded-lg transition-all duration-200 cursor-pointer"
+                onClick={handleLogout}
               >
-                logout
+                Sign out
               </button>
             </div>
           </div>
         ) : (
-          <div className="flex flex-col items-center gap-4">
+          <div className="flex flex-col items-center py-5 h-full">
             <img
               src={catIcon}
               alt="Simba"
-              className="cursor-pointer hover:opacity-80"
+              className="w-9 h-9 cursor-pointer hover:scale-110 transition-transform duration-200"
               onClick={() => setSidebarCollapsed(false)}
             />
           </div>
@@ -236,54 +306,79 @@ export const ChatScreen = ({ setAuthenticated }: ChatScreenProps) => {
       {/* Main chat area */}
       <div className="flex-1 flex flex-col h-screen overflow-hidden">
         {/* Mobile header */}
-        <header className="md:hidden flex flex-row justify-between items-center gap-3 p-4 border-b border-gray-200 bg-white">
+        <header className="md:hidden flex items-center justify-between px-4 py-3 bg-warm-white border-b border-sand-light">
-          <div className="flex flex-row items-center gap-2">
+          <div className="flex items-center gap-2.5">
-            <img src={catIcon} alt="Simba" className="w-10 h-10" />
+            <img src={catIcon} alt="Simba" className="w-8 h-8" />
-            <h1 className="text-xl">asksimba!</h1>
+            <h1 className="font-[family-name:var(--font-display)] text-lg font-bold text-charcoal">
+              asksimba
+            </h1>
           </div>
-          <div className="flex flex-row gap-2">
+          <div className="flex items-center gap-2">
             <button
-              className="p-2 border border-green-400 bg-green-200 hover:bg-green-400 cursor-pointer rounded-md text-sm"
+              className="px-3 py-1.5 text-xs font-medium rounded-lg bg-cream-dark text-charcoal
+                hover:bg-sand-light transition-colors cursor-pointer"
               onClick={() => setShowConversations(!showConversations)}
             >
-              {showConversations ? "hide" : "show"}
+              {showConversations ? "Hide" : "Threads"}
             </button>
             <button
-              className="p-2 border border-red-400 bg-red-200 hover:bg-red-400 cursor-pointer rounded-md text-sm"
+              className="px-3 py-1.5 text-xs font-medium rounded-lg text-warm-gray
-              onClick={() => setAuthenticated(false)}
+                hover:bg-cream-dark transition-colors cursor-pointer"
+              onClick={handleLogout}
             >
-              logout
+              Sign out
             </button>
           </div>
         </header>
 
-        {/* Messages area */}
+        {/* Conversation title bar */}
         {selectedConversation && (
-          <div className="sticky top-0 mx-auto w-full">
+          <div className="bg-warm-white/80 backdrop-blur-sm border-b border-sand-light/50 px-6 py-3">
-            <div className="bg-[#F9F5EB] text-black px-6 w-full py-3">
+            <h2 className="text-sm font-semibold text-charcoal truncate max-w-2xl mx-auto">
-              <h2 className="text-lg font-semibold">
               {selectedConversation.title || "Untitled Conversation"}
             </h2>
           </div>
-          </div>
         )}
-        <div className="flex-1 overflow-y-auto relative px-4 py-6">
-          {/* Floating conversation name */}
 
+        {/* Messages area */}
+        <div className="flex-1 overflow-y-auto px-4 py-6">
           <div className="max-w-2xl mx-auto flex flex-col gap-4">
+            {/* Mobile conversation list */}
             {showConversations && (
-              <div className="md:hidden">
+              <div className="md:hidden mb-2">
                 <ConversationList
                   conversations={conversations}
                   onCreateNewConversation={handleCreateNewConversation}
                   onSelectConversation={handleSelectConversation}
+                  selectedId={selectedConversation?.id}
                 />
               </div>
             )}
+
+            {/* Empty state */}
+            {messages.length === 0 && !isLoading && (
+              <div className="flex flex-col items-center justify-center py-20 gap-4">
+                <div className="relative">
+                  <div className="absolute -inset-4 bg-amber-soft/20 rounded-full blur-2xl" />
+                  <img
+                    src={catIcon}
+                    alt="Simba"
+                    className="relative w-16 h-16 opacity-60"
+                  />
+                </div>
+                <div className="text-center">
+                  <p className="text-warm-gray text-sm">
+                    Ask Simba anything
+                  </p>
+                </div>
+              </div>
+            )}
+
             {messages.map((msg, index) => {
-              if (msg.speaker === "simba") {
+              if (msg.speaker === "tool")
+                return <ToolBubble key={index} text={msg.text} />;
+              if (msg.speaker === "simba")
                 return <AnswerBubble key={index} text={msg.text} />;
-              }
               return <QuestionBubble key={index} text={msg.text} />;
             })}
             {isLoading && <AnswerBubble text="" loading={true} />}
@@ -292,8 +387,8 @@ export const ChatScreen = ({ setAuthenticated }: ChatScreenProps) => {
         </div>
 
         {/* Input area */}
-        <footer className="p-4 bg-[#F9F5EB]">
+        <footer className="border-t border-sand-light/50 bg-warm-white/60 backdrop-blur-sm">
-          <div className="max-w-2xl mx-auto">
+          <div className="max-w-2xl mx-auto px-4 py-4">
             <MessageInput
               query={query}
               handleQueryChange={handleQueryChange}

raggr-frontend/src/components/ConversationList.tsx

@@ -44,12 +44,12 @@ export const ConversationList = ({
   }, []);
 
   return (
-    <div className="bg-indigo-300 rounded-md p-3 sm:p-4 flex flex-col gap-1">
+    <div className="bg-stone-200 rounded-md p-3 sm:p-4 flex flex-col gap-1">
       {conservations.map((conversation) => {
         return (
           <div
             key={conversation.id}
-            className="border-blue-400 bg-indigo-300 hover:bg-indigo-200 cursor-pointer rounded-md p-3 min-h-[44px] flex items-center"
+            className="bg-stone-200 hover:bg-stone-300 cursor-pointer rounded-md p-3 min-h-[44px] flex items-center"
             onClick={() => onSelectConversation(conversation)}
           >
             <p className="text-sm sm:text-base truncate w-full">
@@ -59,7 +59,7 @@ export const ConversationList = ({
         );
       })}
       <div
-        className="border-blue-400 bg-indigo-300 hover:bg-indigo-200 cursor-pointer rounded-md p-3 min-h-[44px] flex items-center"
+        className="bg-stone-200 hover:bg-stone-300 cursor-pointer rounded-md p-3 min-h-[44px] flex items-center"
         onClick={() => onCreateNewConversation()}
       >
         <p className="text-sm sm:text-base"> + Start a new thread</p>

raggr-frontend/src/components/LoginScreen.tsx

@@ -1,6 +1,7 @@
 import { useState, useEffect } from "react";
 import { userService } from "../api/userService";
 import { oidcService } from "../api/oidcService";
+import catIcon from "../assets/cat.png";
 
 type LoginScreenProps = {
   setAuthenticated: (isAuth: boolean) => void;
@@ -76,54 +77,77 @@ export const LoginScreen = ({ setAuthenticated }: LoginScreenProps) => {
   // Show loading state while checking authentication or processing callback
   if (isChecking || isLoggingIn) {
     return (
-      <div className="h-screen bg-opacity-20">
+      <div className="h-screen flex flex-col items-center justify-center bg-cream gap-4">
-        <div className="bg-white/85 h-screen flex items-center justify-center">
+        <img
-          <div className="text-center">
+          src={catIcon}
-            <p className="text-lg sm:text-xl">
+          alt="Simba"
-              {isLoggingIn ? "Logging in..." : "Checking authentication..."}
+          className="w-16 h-16 animate-bounce"
+        />
+        <p className="text-warm-gray font-medium text-lg tracking-wide">
+          {isLoggingIn ? "letting you in..." : "checking credentials..."}
         </p>
       </div>
-        </div>
-      </div>
     );
   }
 
   return (
-    <div className="h-screen bg-opacity-20">
+    <div className="h-screen bg-cream flex items-center justify-center p-4">
-      <div className="bg-white/85 h-screen">
+      {/* Decorative background texture */}
-        <div className="flex flex-row justify-center py-4">
+      <div className="fixed inset-0 opacity-[0.03] pointer-events-none"
-          <div className="flex flex-col gap-4 w-full px-4 sm:w-11/12 sm:max-w-2xl lg:max-w-4xl sm:px-0">
+        style={{
-            <div className="flex flex-col gap-4">
+          backgroundImage: `radial-gradient(circle at 1px 1px, var(--color-charcoal) 1px, transparent 0)`,
-              <div className="flex flex-grow justify-center w-full bg-amber-400 p-2">
+          backgroundSize: '24px 24px'
-                <h1 className="text-base sm:text-xl font-bold text-center">
+        }}
-                  I AM LOOKING FOR A DESIGNER. THIS APP WILL REMAIN UGLY UNTIL A
+      />
-                  DESIGNER COMES.
-                </h1>
-              </div>
-              <header className="flex flex-row justify-center gap-2 grow sticky top-0 z-10 bg-white">
-                <h1 className="text-2xl sm:text-3xl">ask simba!</h1>
-              </header>
 
+      <div className="relative w-full max-w-sm">
+        {/* Cat icon & branding */}
+        <div className="flex flex-col items-center mb-8">
+          <div className="relative mb-4">
+            <div className="absolute -inset-3 bg-amber-soft/40 rounded-full blur-xl" />
+            <img
+              src={catIcon}
+              alt="Simba"
+              className="relative w-20 h-20 drop-shadow-lg"
+            />
+          </div>
+          <h1 className="font-[family-name:var(--font-display)] text-4xl font-bold text-charcoal tracking-tight">
+            asksimba
+          </h1>
+          <p className="text-warm-gray text-sm mt-1.5 tracking-wide">
+            your feline knowledge companion
+          </p>
+        </div>
+
+        {/* Login card */}
+        <div className="bg-warm-white rounded-2xl shadow-lg shadow-sand/40 border border-sand-light/60 p-8">
           {error && (
-                <div className="text-red-600 font-semibold text-sm sm:text-base bg-red-50 p-3 rounded-md">
+            <div className="mb-4 text-sm bg-red-50 text-red-700 p-3 rounded-xl border border-red-200">
               {error}
             </div>
           )}
 
-              <div className="text-center text-sm sm:text-base text-gray-600 py-2">
+          <p className="text-center text-warm-gray text-sm mb-6">
-                Click below to login with Authelia
+            Sign in to start chatting with Simba
-              </div>
+          </p>
-            </div>
 
           <button
-              className="p-3 sm:p-4 min-h-[44px] border border-blue-400 bg-blue-200 hover:bg-blue-400 cursor-pointer rounded-md flex-grow text-sm sm:text-base font-semibold"
+            className="w-full py-3.5 px-4 bg-forest text-white font-semibold rounded-xl
+              hover:bg-forest-light transition-all duration-200
+              active:scale-[0.98] disabled:opacity-50 disabled:cursor-not-allowed
+              shadow-md shadow-forest/20 hover:shadow-lg hover:shadow-forest/30
+              cursor-pointer text-sm tracking-wide"
             onClick={handleOIDCLogin}
             disabled={isLoggingIn}
           >
-              {isLoggingIn ? "Redirecting..." : "Login with Authelia"}
+            {isLoggingIn ? "Redirecting..." : "Sign in with Authelia"}
           </button>
         </div>
-        </div>
+
+        {/* Footer paw prints */}
+        <p className="text-center text-sand mt-6 text-xs tracking-widest select-none">
+          ~ meow ~
+        </p>
       </div>
     </div>
   );

raggr-frontend/src/components/ToolBubble.tsx

@@ -0,0 +1,5 @@
+export const ToolBubble = ({ text }: { text: string }) => (
+  <div className="text-sm text-gray-500 italic px-3 py-1 self-start">
+    {text}
+  </div>
+);

startup.sh

@@ -3,8 +3,17 @@
 echo "Running database migrations..."
 aerich upgrade
 
-echo "Starting reindex process..."
+# Ensure Obsidian vault directory exists
-python main.py "" --reindex
+mkdir -p /app/data/obsidian
 
-echo "Starting Flask application..."
+# Start continuous Obsidian sync if enabled
+if [ "${OBSIDIAN_CONTINUOUS_SYNC}" = "true" ]; then
+    echo "Starting Obsidian continuous sync in background..."
+    ob sync --continuous &
+fi
+
+echo "Starting reindex process in background..."
+python main.py "" --reindex &
+
+echo "Starting application..."
 python app.py

tickets.md

@@ -0,0 +1,189 @@
+# Integration: Twilio API for WhatsApp Interface (Multi-User)
+
+## Overview
+Integrate Twilio's WhatsApp API to allow users to interact with Simba via WhatsApp. This requires multi-user support, linking WhatsApp numbers to existing or new user accounts.
+
+## Tasks
+
+### Phase 1: Infrastructure and Database Changes
+- [x] **[TICKET-001]** Update `User` model to include `whatsapp_number`.
+- [x] **[TICKET-002]** Generate and apply migrations for the database changes.
+
+### Phase 2: Twilio Integration Blueprint
+- [x] **[TICKET-003]** Create a new blueprint for Twilio/WhatsApp webhook.
+- [x] **[TICKET-004]** Implement Twilio signature validation for security.
+    - Decorator enabled on webhook. Set `TWILIO_SIGNATURE_VALIDATION=false` to disable in dev. Set `TWILIO_WEBHOOK_URL` if behind a reverse proxy.
+- [x] **[TICKET-005]** Implement User identification from WhatsApp phone number.
+
+### Phase 3: Core Messaging Logic
+- [x] **[TICKET-006]** Integrate `consult_simba_oracle` with the WhatsApp blueprint.
+- [x] **[TICKET-007]** Implement outgoing WhatsApp message responses.
+- [x] **[TICKET-008]** Handle conversation context in WhatsApp.
+
+### Phase 4: Configuration and Deployment
+- [x] **[TICKET-009]** Add Twilio credentials to environment variables.
+    - Keys: `TWILIO_ACCOUNT_SID`, `TWILIO_AUTH_TOKEN`, `TWILIO_WHATSAPP_NUMBER`.
+- [ ] **[TICKET-010]** Document the Twilio webhook setup in `docs/whatsapp_integration.md`.
+    - Include: Webhook URL format, Twilio Console setup instructions.
+
+### Phase 5: Multi-user & Edge Cases
+- [ ] **[TICKET-011]** Handle first-time users (auto-creation of accounts or invitation system).
+- [ ] **[TICKET-012]** Handle media messages (optional/future: images, audio).
+- [x] **[TICKET-013]** Rate limiting and error handling for Twilio requests.
+
+## Implementation Details
+
+### Twilio Webhook Payload (POST)
+- `SmsMessageSid`, `NumMedia`, `Body`, `From`, `To`, `AccountSid`, etc.
+- We primarily care about `Body` (user message) and `From` (user WhatsApp number).
+
+### Workflow
+1. Twilio receives a message -> POST to `/api/whatsapp/webhook`.
+2. Validate signature.
+3. Identify `User` by `From` number.
+4. If not found, create a new `User` or return an error.
+5. Get/create `Conversation` for this `User`.
+6. Call `consult_simba_oracle` with the query and context.
+7. Return response via TwiML `<Message>` tag.
+
+---
+
+# Integration: Obsidian Bidirectional Data Store
+
+## Overview
+Integrate Obsidian as a bidirectional data store using the [`obsidian-headless`](https://github.com/obsidianmd/obsidian-headless) npm package. SimbaRAG will be able to read/search Obsidian notes for RAG context and write new notes, research summaries, and tasks back to the vault via the LangChain agent.
+
+## Tasks
+
+### Phase 1: Infrastructure
+- [ ] **[OBS-001]** Upgrade Node.js from 20 to 22 in `Dockerfile` (required by obsidian-headless).
+- [ ] **[OBS-002]** Install `obsidian-headless` globally via npm in `Dockerfile`.
+- [ ] **[OBS-003]** Add `obsidian_vault_data` volume and Obsidian env vars to `docker-compose.yml`.
+- [ ] **[OBS-004]** Document Obsidian env vars in `.env.example` (`OBSIDIAN_AUTH_TOKEN`, `OBSIDIAN_VAULT_ID`, `OBSIDIAN_E2E_PASSWORD`, `OBSIDIAN_DEVICE_NAME`, `OBSIDIAN_CONTINUOUS_SYNC`).
+- [ ] **[OBS-005]** Update `startup.sh` to conditionally run `ob sync --continuous` in background when `OBSIDIAN_CONTINUOUS_SYNC=true`.
+
+### Phase 2: Core Service
+- [ ] **[OBS-006]** Create `utils/obsidian_service.py` with `ObsidianService` class.
+    - Vault setup via `ob sync-setup` (async subprocess)
+    - One-time sync via `ob sync`
+    - Sync status via `ob sync-status`
+    - Walk vault directory for `.md` files (skip `.obsidian/`)
+    - Parse Obsidian markdown: YAML frontmatter → metadata, wikilink conversion, embed stripping, tag extraction
+    - Read specific note by relative path
+    - Create new note with frontmatter (auto-adds `created_by: simbarag` + timestamp)
+    - Create task note in configurable tasks folder
+
+### Phase 3: RAG Indexing (Read)
+- [ ] **[OBS-007]** Add `fetch_obsidian_documents()` to `blueprints/rag/logic.py` — uses `ObsidianService` to parse all vault `.md` files into LangChain `Document` objects with `source=obsidian` metadata.
+- [ ] **[OBS-008]** Add `index_obsidian_documents()` to `blueprints/rag/logic.py` — deletes existing `source=obsidian` chunks, splits documents with shared `text_splitter`, embeds into shared `vector_store`.
+- [ ] **[OBS-009]** Add `POST /api/rag/index-obsidian` endpoint (`@admin_required`) to `blueprints/rag/__init__.py`.
+
+### Phase 4: Agent Tools (Read + Write)
+- [ ] **[OBS-010]** Add `obsidian_search_notes` tool to `blueprints/conversation/agents.py` — semantic search via ChromaDB with `where={"source": "obsidian"}` filter.
+- [ ] **[OBS-011]** Add `obsidian_read_note` tool to `blueprints/conversation/agents.py` — reads a specific note by relative path.
+- [ ] **[OBS-012]** Add `obsidian_create_note` tool to `blueprints/conversation/agents.py` — creates a new markdown note in the vault (title, content, folder, tags).
+- [ ] **[OBS-013]** Add `obsidian_create_task` tool to `blueprints/conversation/agents.py` — creates a task note with optional due date.
+- [ ] **[OBS-014]** Register Obsidian tools conditionally (follow YNAB pattern: `obsidian_enabled` flag).
+- [ ] **[OBS-015]** Update system prompt in `blueprints/conversation/__init__.py` with Obsidian tool usage instructions.
+
+### Phase 5: Testing & Verification
+- [ ] **[OBS-016]** Verify Docker image builds with Node.js 22 + obsidian-headless.
+- [ ] **[OBS-017]** Test vault sync: setup → sync → verify files appear in `/app/data/obsidian`.
+- [ ] **[OBS-018]** Test indexing: `POST /api/rag/index-obsidian` → verify chunks in ChromaDB with `source=obsidian`.
+- [ ] **[OBS-019]** Test agent read tools: chat queries trigger `obsidian_search_notes` and `obsidian_read_note`.
+- [ ] **[OBS-020]** Test agent write tools: chat creates notes/tasks → files appear in vault → sync pushes to Obsidian.
+
+## Implementation Details
+
+### Key Files
+- `utils/obsidian_service.py` — new, core service (follows `utils/ynab_service.py` pattern)
+- `blueprints/conversation/agents.py` — add tools (follows YNAB tool pattern at lines 101-279)
+- `blueprints/conversation/__init__.py` — update system prompt (line ~94)
+- `blueprints/rag/logic.py` — add indexing functions (reuse `vector_store`, `text_splitter`)
+- `blueprints/rag/__init__.py` — add index endpoint
+
+### Write-back Model
+Files written to the vault directory are automatically synced to Obsidian Sync by the `ob sync --continuous` background process. No separate push step needed.
+
+### Environment Variables
+| Variable | Required | Description |
+|----------|----------|-------------|
+| `OBSIDIAN_AUTH_TOKEN` | Yes | Auth token for Obsidian Sync (non-interactive) |
+| `OBSIDIAN_VAULT_ID` | Yes | Remote vault ID or name |
+| `OBSIDIAN_E2E_PASSWORD` | If E2EE | End-to-end encryption password |
+| `OBSIDIAN_DEVICE_NAME` | No | Client identifier (default: `simbarag-server`) |
+| `OBSIDIAN_CONTINUOUS_SYNC` | No | Enable background sync (default: `false`) |
+
+---
+
+# Integration: WhatsApp to LangChain Agent Migration
+
+## Overview
+Migrate the WhatsApp blueprint from custom LLM logic to the LangChain agent-based system used by the conversation blueprint. This will provide Tavily web search, YNAB integration, and improved message handling capabilities.
+
+## Tasks
+
+### Phase 1: Import and Setup Changes
+- [x] **[WA-001]** Remove dependency on `main.py`'s `consult_simba_oracle` import in `blueprints/whatsapp/__init__.py`.
+- [x] **[WA-002]** Import `main_agent` from `blueprints.conversation.agents` in `blueprints/whatsapp/__init__.py`.
+- [ ] **[WA-003]** Add import for `query_vector_store` from `blueprints.rag.logic` (if needed for simba_search tool).
+- [x] **[WA-004]** Verify `main_agent` is already initialized as a global variable in `agents.py` (it is at line 295).
+
+### Phase 2: Agent Invocation Adaptation
+- [x] **[WA-005]** Replace `consult_simba_oracle()` call (lines 171-178) with LangChain agent invocation.
+- [x] **[WA-006]** Add system prompt with Simba facts, medical conditions, and recent events from `blueprints/conversation/__init__.py` (lines 55-95).
+- [x] **[WA-007]** Build messages payload with role-based conversation history (last 10 messages).
+- [x] **[WA-008]** Handle agent response extraction: `response.get("messages", [])[-1].content`.
+- [x] **[WA-009]** Keep existing error handling around agent invocation (try/except block).
+
+### Phase 3: Configuration and Logging
+- [x] **[WA-010]** Add YNAB availability logging (check `os.getenv("YNAB_ACCESS_TOKEN")` is not None) in webhook handler.
+- [x] **[WA-011]** Ensure `main_agent` tools include `simba_search`, `web_search`, and optionally YNAB tools (already configured in `agents.py`).
+- [x] **[WA-012]** Verify `simba_search` tool uses `query_vector_store()` which supports `where={"source": "paperless"}` filter (no change needed, works with existing ChromaDB collection).
+
+### Phase 4: Testing Strategy
+- [ ] **[WA-013]** Test Simba queries (e.g., "How much does Simba weigh?") — should use `simba_search` tool.
+- [ ] **[WA-014]** Test general chat queries (e.g., "What's the weather?") — should use LLM directly, no tools.
+- [ ] **[WA-015]** Test web search capability (e.g., "What's the latest cat health research?") — should use `web_search` tool with Tavily.
+- [ ] **[WA-016]** Test YNAB integration if configured (e.g., "How much did I spend on food?") — should use appropriate YNAB tool.
+- [ ] **[WA-017]** Test conversation context preservation (send multiple messages in sequence).
+- [ ] **[WA-018]** Test rate limiting still works after migration.
+- [ ] **[WA-019]** Test user creation and allowlist still function correctly.
+- [ ] **[WA-020]** Test error handling for agent failures (returns "Sorry, I'm having trouble thinking right now. 😿").
+
+### Phase 5: Cleanup and Documentation
+- [ ] **[WA-021]** Optionally remove or deprecate deprecated `main.py` functions: `classify_query()`, `consult_oracle()`, `llm_chat()`, `consult_simba_oracle()` (keep for CLI tool usage).
+- [ ] **[WA-022]** Update code comments in `main.py` to indicate WhatsApp no longer uses these functions.
+- [ ] **[WA-023]** Document the agent-based approach in `docs/whatsapp_integration.md` (if file exists) or create new documentation.
+
+## Implementation Details
+
+### Current WhatsApp Flow
+1. Twilio webhook → `blueprints/whatsapp/__init__.webhook()`
+2. Call `consult_simba_oracle(input, transcript)` from `main.py`
+3. `consult_simba_oracle()` uses custom `QueryGenerator` to classify query
+4. Routes to `consult_oracle()` (ChromaDB) or `llm_chat()` (simple chat)
+5. Returns text response
+
+### Target WhatsApp Flow
+1. Twilio webhook → `blueprints/whatsapp/__init__.webhook()`
+2. Build LangChain messages payload with system prompt and conversation history
+3. Invoke `main_agent.ainvoke({"messages": messages_payload})`
+4. Agent decides when to use tools (simba_search, web_search, YNAB)
+5. Returns text response from last message
+
+### Key Differences
+1. **No manual query classification** — Agent decides based on LLM reasoning
+2. **Tavily web_search** now available for current information
+3. **YNAB integration** available if configured
+4. **System prompt consistency** with conversation blueprint
+5. **Message format** — LangChain messages array vs transcript string
+
+### Environment Variables
+No new environment variables needed. Uses existing:
+- `LLAMA_SERVER_URL` — for LLM model
+- `TAVILY_API_KEY` — for web search
+- `YNAB_ACCESS_TOKEN` — for budget integration (optional)
+
+### Files Modified
+- `blueprints/whatsapp/__init__.py` — Main webhook handler

utils/obsidian_service.py

@@ -0,0 +1,446 @@
+"""Obsidian headless sync service for querying and modifying vaults."""
+
+import os
+import re
+import yaml
+from datetime import datetime
+from pathlib import Path
+from typing import Any, Optional
+from subprocess import run
+
+from dotenv import load_dotenv
+
+# Load environment variables
+load_dotenv()
+
+
+class ObsidianService:
+    """Service for interacting with Obsidian vault via obsidian-headless CLI."""
+
+    def __init__(self):
+        """Initialize Obsidian Sync client."""
+        self.vault_path = os.getenv("OBSIDIAN_VAULT_PATH", "/app/data/obsidian")
+
+        # Create vault path if it doesn't exist
+        Path(self.vault_path).mkdir(parents=True, exist_ok=True)
+
+        # Validate vault has .md files
+        self._validate_vault()
+
+    def _validate_vault(self) -> None:
+        """Validate that vault directory exists and has .md files."""
+        vault_dir = Path(self.vault_path)
+        if not vault_dir.exists():
+            raise ValueError(
+                f"Obsidian vault path '{self.vault_path}' does not exist. "
+                "Please ensure the vault is synced to this location."
+            )
+
+        md_files = list(vault_dir.rglob("*.md"))
+        if not md_files:
+            raise ValueError(
+                f"Vault at '{self.vault_path}' contains no markdown files. "
+                "Please ensure the vault is synced with obsidian-headless."
+            )
+
+    def walk_vault(self) -> list[Path]:
+        """Walk through vault directory and return paths to .md files.
+
+        Returns:
+            List of paths to markdown files, excluding .obsidian directory.
+        """
+        vault_dir = Path(self.vault_path)
+        md_files = []
+
+        # Walk vault, excluding .obsidian directory
+        for md_file in vault_dir.rglob("*.md"):
+            # Skip .obsidian directory and its contents
+            if ".obsidian" in md_file.parts:
+                continue
+            md_files.append(md_file)
+
+        return md_files
+
+    def parse_markdown(self, content: str, filepath: Optional[Path] = None) -> dict[str, Any]:
+        """Parse Obsidian markdown to extract metadata and clean content.
+
+        Args:
+            content: Raw markdown content
+            filepath: Optional file path for context
+
+        Returns:
+            Dictionary containing parsed content:
+                - metadata: Parsed YAML frontmatter (or empty dict if none)
+                - content: Cleaned body content
+                - tags: Extracted tags
+                - wikilinks: List of wikilinks found
+                - embeds: List of embeds found
+        """
+        # Split frontmatter from content
+        frontmatter_pattern = r"^---\n(.*?)\n---"
+        match = re.match(frontmatter_pattern, content, re.DOTALL)
+
+        metadata = {}
+        body_content = content
+
+        if match:
+            frontmatter = match.group(1)
+            body_content = content[match.end():].strip()
+            try:
+                metadata = yaml.safe_load(frontmatter) or {}
+            except yaml.YAMLError:
+                # Invalid YAML, treat as empty metadata
+                metadata = {}
+
+        # Extract tags (#tag format)
+        tags = re.findall(r"#(\w+)", content)
+        tags = [tag for tag in tags if tag]  # Remove empty strings
+
+        # Extract wikilinks [[wiki link]]
+        wikilinks = re.findall(r"\[\[([^\]]+)\]\]", content)
+
+        # Extract embeds [[!embed]] or [[!embed:file]]
+        embeds = re.findall(r"\[\[!(.*?)\]\]", content)
+        embeds = [e.split(":")[0].strip() if ":" in e else e.strip() for e in embeds]
+
+        # Clean body content
+        # Remove wikilinks [[...]] and embeds [[!...]]
+        cleaned_content = re.sub(r"\[\[.*?\]\]", "", body_content)
+        cleaned_content = re.sub(r"\n{3,}", "\n\n", cleaned_content).strip()
+
+        return {
+            "metadata": metadata,
+            "content": cleaned_content,
+            "tags": tags,
+            "wikilinks": wikilinks,
+            "embeds": embeds,
+            "filepath": str(filepath) if filepath else None,
+        }
+
+    def read_note(self, relative_path: str) -> dict[str, Any]:
+        """Read a specific note from the vault.
+
+        Args:
+            relative_path: Path to note relative to vault root (e.g., "My Notes/simba.md")
+
+        Returns:
+            Dictionary containing parsed note content and metadata.
+        """
+        vault_dir = Path(self.vault_path)
+        note_path = vault_dir / relative_path
+
+        if not note_path.exists():
+            raise FileNotFoundError(f"Note not found at '{relative_path}'")
+
+        with open(note_path, "r", encoding="utf-8") as f:
+            content = f.read()
+
+        parsed = self.parse_markdown(content, note_path)
+
+        return {
+            "content": parsed,
+            "path": relative_path,
+            "full_path": str(note_path),
+        }
+
+    def create_note(
+        self,
+        title: str,
+        content: str,
+        folder: str = "notes",
+        tags: Optional[list[str]] = None,
+        frontmatter: Optional[dict[str, Any]] = None,
+    ) -> str:
+        """Create a new note in the vault.
+
+        Args:
+            title: Note title (will be used as filename)
+            content: Note body content
+            folder: Folder path (default: "notes")
+            tags: List of tags to add
+            frontmatter: Optional custom frontmatter to merge with defaults
+
+        Returns:
+            Path to created note (relative to vault root).
+        """
+        vault_dir = Path(self.vault_path)
+        note_folder = vault_dir / folder
+        note_folder.mkdir(parents=True, exist_ok=True)
+
+        # Sanitize title for filename
+        safe_title = re.sub(r"[^a-z0-9-_]", "-", title.lower().strip())
+        safe_title = re.sub(r"-+", "-", safe_title).strip("-")
+
+        note_path = note_folder / f"{safe_title}.md"
+
+        # Build frontmatter
+        default_frontmatter = {
+            "created_by": "simbarag",
+            "created_at": datetime.now().isoformat(),
+        }
+
+        if frontmatter:
+            default_frontmatter.update(frontmatter)
+
+        # Add tags to frontmatter if provided
+        if tags:
+            default_frontmatter.setdefault("tags", []).extend(tags)
+
+        # Write note
+        frontmatter_yaml = yaml.dump(default_frontmatter, allow_unicode=True, default_flow_style=False)
+        full_content = f"---\n{frontmatter_yaml}---\n\n{content}"
+
+        with open(note_path, "w", encoding="utf-8") as f:
+            f.write(full_content)
+
+        return f"{folder}/{safe_title}.md"
+
+    def create_task(
+        self,
+        title: str,
+        content: str = "",
+        folder: str = "tasks",
+        due_date: Optional[str] = None,
+        tags: Optional[list[str]] = None,
+    ) -> str:
+        """Create a task note in the vault.
+
+        Args:
+            title: Task title
+            content: Task description
+            folder: Folder to place task (default: "tasks")
+            due_date: Optional due date in YYYY-MM-DD format
+            tags: Optional list of tags to add
+
+        Returns:
+            Path to created task note (relative to vault root).
+        """
+        task_content = f"# {title}\n\n{content}"
+
+        # Add checkboxes if content is empty (simple task)
+        if not content.strip():
+            task_content += "\n- [ ]"
+
+        # Add due date if provided
+        if due_date:
+            task_content += f"\n\n**Due**: {due_date}"
+
+        # Add tags if provided
+        if tags:
+            task_content += "\n\n" + " ".join([f"#{tag}" for tag in tags])
+
+        return self.create_note(
+            title=title,
+            content=task_content,
+            folder=folder,
+            tags=tags,
+        )
+
+    def get_daily_note_path(self, date: Optional[datetime] = None) -> str:
+        """Return the relative vault path for a daily note.
+
+        Args:
+            date: Date for the note (defaults to today)
+
+        Returns:
+            Relative path like "journal/2026/2026-03-03.md"
+        """
+        if date is None:
+            date = datetime.now()
+        return f"journal/{date.strftime('%Y')}/{date.strftime('%Y-%m-%d')}.md"
+
+    def get_daily_note(self, date: Optional[datetime] = None) -> dict[str, Any]:
+        """Read a daily note from the vault.
+
+        Args:
+            date: Date for the note (defaults to today)
+
+        Returns:
+            Dictionary with found status, path, raw content, and date string.
+        """
+        if date is None:
+            date = datetime.now()
+        relative_path = self.get_daily_note_path(date)
+        note_path = Path(self.vault_path) / relative_path
+
+        if not note_path.exists():
+            return {"found": False, "path": relative_path, "content": None, "date": date.strftime("%Y-%m-%d")}
+
+        with open(note_path, "r", encoding="utf-8") as f:
+            content = f.read()
+
+        return {"found": True, "path": relative_path, "content": content, "date": date.strftime("%Y-%m-%d")}
+
+    def get_daily_tasks(self, date: Optional[datetime] = None) -> dict[str, Any]:
+        """Extract tasks from a daily note's tasks section.
+
+        Args:
+            date: Date for the note (defaults to today)
+
+        Returns:
+            Dictionary with tasks list (each has "text" and "done" keys) and metadata.
+        """
+        if date is None:
+            date = datetime.now()
+        note = self.get_daily_note(date)
+        if not note["found"]:
+            return {"found": False, "tasks": [], "date": note["date"], "path": note["path"]}
+
+        tasks = []
+        in_tasks = False
+        for line in note["content"].split("\n"):
+            if re.match(r"^###\s+tasks\s*$", line, re.IGNORECASE):
+                in_tasks = True
+                continue
+            if in_tasks and re.match(r"^#{1,3}\s", line):
+                break
+            if in_tasks:
+                done_match = re.match(r"^- \[x\] (.+)$", line, re.IGNORECASE)
+                todo_match = re.match(r"^- \[ \] (.+)$", line)
+                if done_match:
+                    tasks.append({"text": done_match.group(1), "done": True})
+                elif todo_match:
+                    tasks.append({"text": todo_match.group(1), "done": False})
+
+        return {"found": True, "tasks": tasks, "date": note["date"], "path": note["path"]}
+
+    def add_task_to_daily_note(self, task_text: str, date: Optional[datetime] = None) -> dict[str, Any]:
+        """Add a task checkbox to a daily note, creating the note if needed.
+
+        Args:
+            task_text: The task description text
+            date: Date for the note (defaults to today)
+
+        Returns:
+            Dictionary with success status, path, and whether note was created.
+        """
+        if date is None:
+            date = datetime.now()
+        relative_path = self.get_daily_note_path(date)
+        note_path = Path(self.vault_path) / relative_path
+
+        if not note_path.exists():
+            note_path.parent.mkdir(parents=True, exist_ok=True)
+            content = (
+                f"---\nmodified: {datetime.now().isoformat()}\n---\n"
+                f"###  tasks\n\n- [ ] {task_text}\n\n### log\n"
+            )
+            with open(note_path, "w", encoding="utf-8") as f:
+                f.write(content)
+            return {"success": True, "created_note": True, "path": relative_path}
+
+        with open(note_path, "r", encoding="utf-8") as f:
+            content = f.read()
+
+        # Insert before ### log if present, otherwise append before end
+        log_match = re.search(r"\n(### log)", content, re.IGNORECASE)
+        if log_match:
+            insert_pos = log_match.start()
+            content = content[:insert_pos] + f"\n- [ ] {task_text}" + content[insert_pos:]
+        else:
+            content = content.rstrip() + f"\n- [ ] {task_text}\n"
+
+        with open(note_path, "w", encoding="utf-8") as f:
+            f.write(content)
+
+        return {"success": True, "created_note": False, "path": relative_path}
+
+    def complete_task_in_daily_note(self, task_text: str, date: Optional[datetime] = None) -> dict[str, Any]:
+        """Mark a task as complete in a daily note by matching task text.
+
+        Searches for a task matching the given text (exact or partial) and
+        replaces `- [ ]` with `- [x]`.
+
+        Args:
+            task_text: The task text to search for (exact or partial match)
+            date: Date for the note (defaults to today)
+
+        Returns:
+            Dictionary with success status, matched task text, and path.
+        """
+        if date is None:
+            date = datetime.now()
+        relative_path = self.get_daily_note_path(date)
+        note_path = Path(self.vault_path) / relative_path
+
+        if not note_path.exists():
+            return {"success": False, "error": "Note not found", "path": relative_path}
+
+        with open(note_path, "r", encoding="utf-8") as f:
+            content = f.read()
+
+        # Try exact match first, then partial
+        exact = f"- [ ] {task_text}"
+        if exact in content:
+            content = content.replace(exact, f"- [x] {task_text}", 1)
+        else:
+            match = re.search(r"- \[ \] .*" + re.escape(task_text) + r".*", content, re.IGNORECASE)
+            if not match:
+                return {"success": False, "error": f"Task '{task_text}' not found", "path": relative_path}
+            completed = match.group(0).replace("- [ ]", "- [x]", 1)
+            content = content.replace(match.group(0), completed, 1)
+            task_text = match.group(0).replace("- [ ] ", "")
+
+        with open(note_path, "w", encoding="utf-8") as f:
+            f.write(content)
+
+        return {"success": True, "completed_task": task_text, "path": relative_path}
+
+    def sync_vault(self) -> dict[str, Any]:
+        """Trigger a one-time sync of the vault.
+
+        Returns:
+            Dictionary containing sync result and output.
+        """
+        try:
+            result = run(
+                ["ob", "sync"],
+                capture_output=True,
+                text=True,
+                timeout=300,
+            )
+
+            if result.returncode != 0:
+                return {
+                    "success": False,
+                    "error": result.stderr or "Sync failed",
+                    "stdout": result.stdout,
+                }
+
+            return {
+                "success": True,
+                "message": "Vault synced successfully",
+                "stdout": result.stdout,
+            }
+
+        except Exception as e:
+            return {
+                "success": False,
+                "error": str(e),
+            }
+
+    def sync_status(self) -> dict[str, Any]:
+        """Check sync status of the vault.
+
+        Returns:
+            Dictionary containing sync status information.
+        """
+        try:
+            result = run(
+                ["ob", "sync-status"],
+                capture_output=True,
+                text=True,
+                timeout=60,
+            )
+
+            return {
+                "success": True,
+                "output": result.stdout,
+                "stderr": result.stderr,
+            }
+
+        except Exception as e:
+            return {
+                "success": False,
+                "error": str(e),
+            }

utils/ynab_service.py

@@ -0,0 +1,342 @@
+"""YNAB API service for querying budget data."""
+
+import os
+from datetime import datetime, timedelta
+from typing import Any, Optional
+
+from dotenv import load_dotenv
+import ynab
+
+# Load environment variables
+load_dotenv()
+
+
+class YNABService:
+    """Service for interacting with YNAB API."""
+
+    def __init__(self):
+        """Initialize YNAB API client."""
+        self.access_token = os.getenv("YNAB_ACCESS_TOKEN", "")
+        self.budget_id = os.getenv("YNAB_BUDGET_ID", "")
+
+        if not self.access_token:
+            raise ValueError("YNAB_ACCESS_TOKEN environment variable is required")
+
+        # Configure API client
+        configuration = ynab.Configuration(access_token=self.access_token)
+        self.api_client = ynab.ApiClient(configuration)
+
+        # Initialize API endpoints
+        self.budgets_api = ynab.BudgetsApi(self.api_client)
+        self.transactions_api = ynab.TransactionsApi(self.api_client)
+        self.months_api = ynab.MonthsApi(self.api_client)
+        self.categories_api = ynab.CategoriesApi(self.api_client)
+
+        # Get budget ID if not provided
+        if not self.budget_id:
+            budgets_response = self.budgets_api.get_budgets()
+            if budgets_response.data and budgets_response.data.budgets:
+                self.budget_id = budgets_response.data.budgets[0].id
+            else:
+                raise ValueError("No YNAB budgets found")
+
+    def get_budget_summary(self) -> dict[str, Any]:
+        """Get overall budget summary and health status.
+
+        Returns:
+            Dictionary containing budget summary with to-be-budgeted amount,
+            total budgeted, total activity, and overall budget health.
+        """
+        budget_response = self.budgets_api.get_budget_by_id(self.budget_id)
+        budget_data = budget_response.data.budget
+
+        # Calculate totals from categories
+        to_be_budgeted = (
+            budget_data.months[0].to_be_budgeted / 1000 if budget_data.months else 0
+        )
+
+        total_budgeted = 0
+        total_activity = 0
+        total_available = 0
+
+        for category_group in budget_data.category_groups or []:
+            if category_group.deleted or category_group.hidden:
+                continue
+            for category in category_group.categories or []:
+                if category.deleted or category.hidden:
+                    continue
+                total_budgeted += category.budgeted / 1000
+                total_activity += category.activity / 1000
+                total_available += category.balance / 1000
+
+        return {
+            "budget_name": budget_data.name,
+            "to_be_budgeted": round(to_be_budgeted, 2),
+            "total_budgeted": round(total_budgeted, 2),
+            "total_activity": round(total_activity, 2),
+            "total_available": round(total_available, 2),
+            "currency_format": budget_data.currency_format.iso_code
+            if budget_data.currency_format
+            else "USD",
+            "summary": f"Budget '{budget_data.name}' has ${abs(to_be_budgeted):.2f} {'to be budgeted' if to_be_budgeted > 0 else 'overbudgeted'}. "
+            f"Total budgeted: ${total_budgeted:.2f}, Total spent: ${abs(total_activity):.2f}, "
+            f"Total available: ${total_available:.2f}.",
+        }
+
+    def get_transactions(
+        self,
+        start_date: Optional[str] = None,
+        end_date: Optional[str] = None,
+        category_name: Optional[str] = None,
+        payee_name: Optional[str] = None,
+        limit: int = 50,
+    ) -> dict[str, Any]:
+        """Get transactions filtered by date range, category, or payee.
+
+        Args:
+            start_date: Start date in YYYY-MM-DD format (defaults to 30 days ago)
+            end_date: End date in YYYY-MM-DD format (defaults to today)
+            category_name: Filter by category name (case-insensitive partial match)
+            payee_name: Filter by payee name (case-insensitive partial match)
+            limit: Maximum number of transactions to return (default 50)
+
+        Returns:
+            Dictionary containing matching transactions and summary statistics.
+        """
+        # Set default date range if not provided
+        if not start_date:
+            start_date = (datetime.now() - timedelta(days=30)).strftime("%Y-%m-%d")
+        if not end_date:
+            end_date = datetime.now().strftime("%Y-%m-%d")
+
+        # Get transactions
+        transactions_response = self.transactions_api.get_transactions(
+            self.budget_id, since_date=start_date
+        )
+
+        transactions = transactions_response.data.transactions or []
+
+        # Filter by date range, category, and payee
+        filtered_transactions = []
+        total_amount = 0
+
+        for txn in transactions:
+            # Skip if deleted or before start date or after end date
+            if txn.deleted:
+                continue
+            txn_date = str(txn.date)
+            if txn_date < start_date or txn_date > end_date:
+                continue
+
+            # Filter by category if specified
+            if category_name and txn.category_name:
+                if category_name.lower() not in txn.category_name.lower():
+                    continue
+
+            # Filter by payee if specified
+            if payee_name and txn.payee_name:
+                if payee_name.lower() not in txn.payee_name.lower():
+                    continue
+
+            amount = txn.amount / 1000  # Convert milliunits to dollars
+            filtered_transactions.append(
+                {
+                    "date": txn_date,
+                    "payee": txn.payee_name,
+                    "category": txn.category_name,
+                    "memo": txn.memo,
+                    "amount": round(amount, 2),
+                    "approved": txn.approved,
+                }
+            )
+            total_amount += amount
+
+        # Sort by date (most recent first) and limit
+        filtered_transactions.sort(key=lambda x: x["date"], reverse=True)
+        filtered_transactions = filtered_transactions[:limit]
+
+        return {
+            "transactions": filtered_transactions,
+            "count": len(filtered_transactions),
+            "total_amount": round(total_amount, 2),
+            "start_date": start_date,
+            "end_date": end_date,
+            "filters": {"category": category_name, "payee": payee_name},
+        }
+
+    def get_category_spending(self, month: Optional[str] = None) -> dict[str, Any]:
+        """Get spending breakdown by category for a specific month.
+
+        Args:
+            month: Month in YYYY-MM format (defaults to current month)
+
+        Returns:
+            Dictionary containing spending by category and summary.
+        """
+        if not month:
+            month = datetime.now().strftime("%Y-%m-01")
+        else:
+            # Ensure format is YYYY-MM-01
+            if len(month) == 7:  # YYYY-MM
+                month = f"{month}-01"
+
+        # Get budget month
+        month_response = self.months_api.get_budget_month(self.budget_id, month)
+
+        month_data = month_response.data.month
+
+        categories_spending = []
+        total_budgeted = 0
+        total_activity = 0
+        total_available = 0
+        overspent_categories = []
+
+        for category in month_data.categories or []:
+            if category.deleted or category.hidden:
+                continue
+
+            budgeted = category.budgeted / 1000
+            activity = category.activity / 1000
+            available = category.balance / 1000
+
+            total_budgeted += budgeted
+            total_activity += activity
+            total_available += available
+
+            # Track overspent categories
+            if available < 0:
+                overspent_categories.append(
+                    {
+                        "name": category.name,
+                        "budgeted": round(budgeted, 2),
+                        "spent": round(abs(activity), 2),
+                        "overspent_by": round(abs(available), 2),
+                    }
+                )
+
+            # Only include categories with activity
+            if activity != 0:
+                categories_spending.append(
+                    {
+                        "category": category.name,
+                        "budgeted": round(budgeted, 2),
+                        "activity": round(activity, 2),
+                        "available": round(available, 2),
+                        "goal_type": category.goal_type
+                        if hasattr(category, "goal_type")
+                        else None,
+                    }
+                )
+
+        # Sort by absolute activity (highest spending first)
+        categories_spending.sort(key=lambda x: abs(x["activity"]), reverse=True)
+
+        return {
+            "month": month[:7],  # Return YYYY-MM format
+            "categories": categories_spending,
+            "total_budgeted": round(total_budgeted, 2),
+            "total_spent": round(abs(total_activity), 2),
+            "total_available": round(total_available, 2),
+            "overspent_categories": overspent_categories,
+            "to_be_budgeted": round(month_data.to_be_budgeted / 1000, 2)
+            if month_data.to_be_budgeted
+            else 0,
+        }
+
+    def get_spending_insights(self, months_back: int = 3) -> dict[str, Any]:
+        """Generate insights about spending patterns and budget health.
+
+        Args:
+            months_back: Number of months to analyze (default 3)
+
+        Returns:
+            Dictionary containing spending insights, trends, and recommendations.
+        """
+        current_month = datetime.now()
+        monthly_data = []
+
+        # Collect data for the last N months
+        for i in range(months_back):
+            month_date = current_month - timedelta(days=30 * i)
+            month_str = month_date.strftime("%Y-%m-01")
+
+            try:
+                month_spending = self.get_category_spending(month_str)
+                monthly_data.append(
+                    {
+                        "month": month_str[:7],
+                        "total_spent": month_spending["total_spent"],
+                        "total_budgeted": month_spending["total_budgeted"],
+                        "overspent_categories": month_spending["overspent_categories"],
+                    }
+                )
+            except Exception:
+                # Skip months that don't have data yet
+                continue
+
+        if not monthly_data:
+            return {"error": "No spending data available for analysis"}
+
+        # Calculate average spending
+        avg_spending = sum(m["total_spent"] for m in monthly_data) / len(monthly_data)
+        current_spending = monthly_data[0]["total_spent"] if monthly_data else 0
+
+        # Identify spending trend
+        if len(monthly_data) >= 2:
+            recent_avg = sum(m["total_spent"] for m in monthly_data[:2]) / 2
+            older_avg = sum(m["total_spent"] for m in monthly_data[1:]) / (
+                len(monthly_data) - 1
+            )
+            trend = (
+                "increasing"
+                if recent_avg > older_avg * 1.1
+                else "decreasing"
+                if recent_avg < older_avg * 0.9
+                else "stable"
+            )
+        else:
+            trend = "insufficient data"
+
+        # Find frequently overspent categories
+        overspent_frequency = {}
+        for month in monthly_data:
+            for cat in month["overspent_categories"]:
+                cat_name = cat["name"]
+                if cat_name not in overspent_frequency:
+                    overspent_frequency[cat_name] = 0
+                overspent_frequency[cat_name] += 1
+
+        frequently_overspent = [
+            {"category": cat, "months_overspent": count}
+            for cat, count in overspent_frequency.items()
+            if count > 1
+        ]
+        frequently_overspent.sort(key=lambda x: x["months_overspent"], reverse=True)
+
+        # Generate recommendations
+        recommendations = []
+        if current_spending > avg_spending * 1.2:
+            recommendations.append(
+                f"Current month spending (${current_spending:.2f}) is significantly higher than your {months_back}-month average (${avg_spending:.2f})"
+            )
+
+        if frequently_overspent:
+            top_overspent = frequently_overspent[0]
+            recommendations.append(
+                f"'{top_overspent['category']}' has been overspent in {top_overspent['months_overspent']} of the last {months_back} months"
+            )
+
+        if trend == "increasing":
+            recommendations.append(
+                "Your spending trend is increasing. Consider reviewing your budget allocations."
+            )
+
+        return {
+            "analysis_period": f"Last {months_back} months",
+            "average_monthly_spending": round(avg_spending, 2),
+            "current_month_spending": round(current_spending, 2),
+            "spending_trend": trend,
+            "frequently_overspent_categories": frequently_overspent,
+            "recommendations": recommendations,
+            "monthly_breakdown": monthly_data,
+        }

uv.lock

@@ -103,6 +103,18 @@ wheels = [
     { url = "https://files.pythonhosted.org/packages/b4/63/278a98c715ae467624eafe375542d8ba9b4383a016df8fdefe0ae28382a7/aiohttp-3.13.3-cp314-cp314t-win_amd64.whl", hash = "sha256:44531a36aa2264a1860089ffd4dce7baf875ee5a6079d5fb42e261c704ef7344", size = 499694, upload-time = "2026-01-03T17:32:24.546Z" },
 ]
 
+[[package]]
+name = "aiohttp-retry"
+version = "2.9.1"
+source = { registry = "https://pypi.org/simple" }
+dependencies = [
+    { name = "aiohttp" },
+]
+sdist = { url = "https://files.pythonhosted.org/packages/9d/61/ebda4d8e3d8cfa1fd3db0fb428db2dd7461d5742cea35178277ad180b033/aiohttp_retry-2.9.1.tar.gz", hash = "sha256:8eb75e904ed4ee5c2ec242fefe85bf04240f685391c4879d8f541d6028ff01f1", size = 13608, upload-time = "2024-11-06T10:44:54.574Z" }
+wheels = [
+    { url = "https://files.pythonhosted.org/packages/1a/99/84ba7273339d0f3dfa57901b846489d2e5c2cd731470167757f1935fffbd/aiohttp_retry-2.9.1-py3-none-any.whl", hash = "sha256:66d2759d1921838256a05a3f80ad7e724936f083e35be5abb5e16eed6be6dc54", size = 9981, upload-time = "2024-11-06T10:44:52.917Z" },
+]
+
 [[package]]
 name = "aiosignal"
 version = "1.4.0"
@@ -1281,19 +1293,6 @@ wheels = [
     { url = "https://files.pythonhosted.org/packages/83/bd/9df897cbc98290bf71140104ee5b9777cf5291afb80333aa7da5a497339b/langchain_core-1.2.5-py3-none-any.whl", hash = "sha256:3255944ef4e21b2551facb319bfc426057a40247c0a05de5bd6f2fc021fbfa34", size = 484851, upload-time = "2025-12-22T23:45:30.525Z" },
 ]
 
-[[package]]
-name = "langchain-ollama"
-version = "1.0.1"
-source = { registry = "https://pypi.org/simple" }
-dependencies = [
-    { name = "langchain-core" },
-    { name = "ollama" },
-]
-sdist = { url = "https://files.pythonhosted.org/packages/73/51/72cd04d74278f3575f921084f34280e2f837211dc008c9671c268c578afe/langchain_ollama-1.0.1.tar.gz", hash = "sha256:e37880c2f41cdb0895e863b1cfd0c2c840a117868b3f32e44fef42569e367443", size = 153850, upload-time = "2025-12-12T21:48:28.68Z" }
-wheels = [
-    { url = "https://files.pythonhosted.org/packages/e3/46/f2907da16dc5a5a6c679f83b7de21176178afad8d2ca635a581429580ef6/langchain_ollama-1.0.1-py3-none-any.whl", hash = "sha256:37eb939a4718a0255fe31e19fbb0def044746c717b01b97d397606ebc3e9b440", size = 29207, upload-time = "2025-12-12T21:48:27.832Z" },
-]
-
 [[package]]
 name = "langchain-openai"
 version = "1.1.6"
@@ -1715,15 +1714,15 @@ wheels = [
 
 [[package]]
 name = "ollama"
-version = "0.6.0"
+version = "0.6.1"
 source = { registry = "https://pypi.org/simple" }
 dependencies = [
     { name = "httpx" },
     { name = "pydantic" },
 ]
-sdist = { url = "https://files.pythonhosted.org/packages/d6/47/f9ee32467fe92744474a8c72e138113f3b529fc266eea76abfdec9a33f3b/ollama-0.6.0.tar.gz", hash = "sha256:da2b2d846b5944cfbcee1ca1e6ee0585f6c9d45a2fe9467cbcd096a37383da2f", size = 50811, upload-time = "2025-09-24T22:46:02.417Z" }
+sdist = { url = "https://files.pythonhosted.org/packages/9d/5a/652dac4b7affc2b37b95386f8ae78f22808af09d720689e3d7a86b6ed98e/ollama-0.6.1.tar.gz", hash = "sha256:478c67546836430034b415ed64fa890fd3d1ff91781a9d548b3325274e69d7c6", size = 51620, upload-time = "2025-11-13T23:02:17.416Z" }
 wheels = [
-    { url = "https://files.pythonhosted.org/packages/b5/c1/edc9f41b425ca40b26b7c104c5f6841a4537bb2552bfa6ca66e81405bb95/ollama-0.6.0-py3-none-any.whl", hash = "sha256:534511b3ccea2dff419ae06c3b58d7f217c55be7897c8ce5868dfb6b219cf7a0", size = 14130, upload-time = "2025-09-24T22:46:01.19Z" },
+    { url = "https://files.pythonhosted.org/packages/47/4f/4a617ee93d8208d2bcf26b2d8b9402ceaed03e3853c754940e2290fed063/ollama-0.6.1-py3-none-any.whl", hash = "sha256:fc4c984b345735c5486faeee67d8a265214a31cbb828167782dc642ce0a2bf8c", size = 14354, upload-time = "2025-11-13T23:02:16.292Z" },
 ]
 
 [[package]]
@@ -2534,7 +2533,6 @@ dependencies = [
     { name = "langchain" },
     { name = "langchain-chroma" },
     { name = "langchain-community" },
-    { name = "langchain-ollama" },
     { name = "langchain-openai" },
     { name = "ollama" },
     { name = "openai" },
@@ -2551,6 +2549,8 @@ dependencies = [
     { name = "tomlkit" },
     { name = "tortoise-orm" },
     { name = "tortoise-orm-stubs" },
+    { name = "twilio" },
+    { name = "ynab" },
 ]
 
 [package.metadata]
@@ -2569,9 +2569,8 @@ requires-dist = [
     { name = "langchain", specifier = ">=1.2.0" },
     { name = "langchain-chroma", specifier = ">=1.0.0" },
     { name = "langchain-community", specifier = ">=0.4.1" },
-    { name = "langchain-ollama", specifier = ">=1.0.1" },
     { name = "langchain-openai", specifier = ">=1.1.6" },
-    { name = "ollama", specifier = ">=0.6.0" },
+    { name = "ollama", specifier = ">=0.6.1" },
     { name = "openai", specifier = ">=2.0.1" },
     { name = "pillow", specifier = ">=10.0.0" },
     { name = "pillow-heif", specifier = ">=1.1.1" },
@@ -2584,8 +2583,10 @@ requires-dist = [
     { name = "quart-jwt-extended", specifier = ">=0.1.0" },
     { name = "tavily-python", specifier = ">=0.7.17" },
     { name = "tomlkit", specifier = ">=0.13.3" },
-    { name = "tortoise-orm", specifier = ">=0.25.1" },
+    { name = "tortoise-orm", specifier = ">=0.25.1,<1.0.0" },
     { name = "tortoise-orm-stubs", specifier = ">=1.0.2" },
+    { name = "twilio", specifier = ">=9.10.2" },
+    { name = "ynab", specifier = ">=1.3.0" },
 ]
 
 [[package]]
@@ -3000,6 +3001,21 @@ wheels = [
     { url = "https://files.pythonhosted.org/packages/d0/30/dc54f88dd4a2b5dc8a0279bdd7270e735851848b762aeb1c1184ed1f6b14/tqdm-4.67.1-py3-none-any.whl", hash = "sha256:26445eca388f82e72884e0d580d5464cd801a3ea01e63e5601bdff9ba6a48de2", size = 78540, upload-time = "2024-11-24T20:12:19.698Z" },
 ]
 
+[[package]]
+name = "twilio"
+version = "9.10.2"
+source = { registry = "https://pypi.org/simple" }
+dependencies = [
+    { name = "aiohttp" },
+    { name = "aiohttp-retry" },
+    { name = "pyjwt" },
+    { name = "requests" },
+]
+sdist = { url = "https://files.pythonhosted.org/packages/1c/a1/44cd8604eb69b1c5e7c0f07f0e4305b1884a3b75e23eb8d89350fe7bb982/twilio-9.10.2.tar.gz", hash = "sha256:f17d778870a7419a7278d5747b0e80a1c89e6f5ab14acf5456a004f8f2016bfa", size = 1618748, upload-time = "2026-02-18T04:40:44.279Z" }
+wheels = [
+    { url = "https://files.pythonhosted.org/packages/6c/ac/e1937f70544075f896bfcd6b23fa7c15cad945e4598bcfa7017b7c120ad8/twilio-9.10.2-py2.py3-none-any.whl", hash = "sha256:8722bb59bacf31fab5725d6f5d3fac2224265c669d38f653f53179165533da43", size = 2256481, upload-time = "2026-02-18T04:40:42.226Z" },
+]
+
 [[package]]
 name = "typer"
 version = "0.19.2"
@@ -3385,6 +3401,22 @@ wheels = [
     { url = "https://files.pythonhosted.org/packages/73/ae/b48f95715333080afb75a4504487cbe142cae1268afc482d06692d605ae6/yarl-1.22.0-py3-none-any.whl", hash = "sha256:1380560bdba02b6b6c90de54133c81c9f2a453dee9912fe58c1dcced1edb7cff", size = 46814, upload-time = "2025-10-06T14:12:53.872Z" },
 ]
 
+[[package]]
+name = "ynab"
+version = "1.9.0"
+source = { registry = "https://pypi.org/simple" }
+dependencies = [
+    { name = "certifi" },
+    { name = "pydantic" },
+    { name = "python-dateutil" },
+    { name = "typing-extensions" },
+    { name = "urllib3" },
+]
+sdist = { url = "https://files.pythonhosted.org/packages/9a/3e/36599ae876db3e1d32e393ab0934547df75bab70373c14ca5805246f99bc/ynab-1.9.0.tar.gz", hash = "sha256:fa50bdff641b3a273661e9f6e8a210f5ad98991a998dc09dec0a8122d734d1c6", size = 64898, upload-time = "2025-10-06T19:14:32.707Z" }
+wheels = [
+    { url = "https://files.pythonhosted.org/packages/b2/9c/0ccd11bcdf7522fcb2823fcd7ffbb48e3164d72caaf3f920c7b068347175/ynab-1.9.0-py3-none-any.whl", hash = "sha256:72ac0219605b4280149684ecd0fec3bd75d938772d65cdeea9b3e66a1b2f470d", size = 208674, upload-time = "2025-10-06T19:14:31.719Z" },
+]
+
 [[package]]
 name = "zipp"
 version = "3.23.0"