Notion AI Bot

A modern, async Python bot that connects Notion to OpenAI's ChatGPT, automating Q&A, research, and documentation directly in your Notion workspace.

---

Features

• Seamless Notion ↔️ ChatGPT integration
• Async worker: Fast, scalable, and configurable polling
• Web dashboard: Monitor status via FastAPI (app.py)
• Automatic code extraction: Stores code blocks separately
• Contextual memory: Remembers previous prompts for better answers
• Persistent storage: Uses SQLite for all prompt/response history
• Easy search & reset: CLI tools to view, search, or reset memory
• Auto-reset on inactivity: Memory can be reset automatically after a configurable period of inactivity

---

Quick Start

1. Clone the repo
2. Set up your Notion database (see setup guide)
3. Create a .env file with your API keys (see below)
4. Install dependencies

   pip install -r requirements.txt

5. Run the worker

   python main.py

6. (Optional) Run the web dashboard

   python app.py

---

Environment Variables

• NOTION_DB_ID — Your Notion database ID
• NOTION_API_KEY — Your Notion integration token
• OPENAI_API_KEY — Your OpenAI API key
• FAST_MODE — (optional) Set to 1 for instant response
• CONTEXT_WINDOW — (optional) Number of previous prompts to use as context (default: 5)
• INACTIVITY_RESET_HOURS — (optional) Number of hours of inactivity before memory is automatically reset (default: 24)

---

Usage

• Add prompts to your Notion database (Status: Pending)
• The worker processes them and writes responses back to Notion
• Monitor via the web dashboard (optional)
• View memory:

  python main.py memory

• Search memory:

  python main.py search <query>

• Reset memory:

  python main.py reset

---

Running CLI Commands in Railway (Locally)

You can use the Railway CLI to link your local project to your Railway service and run commands in the Railway environment.

1. Install the Railway CLI

npm install -g railway

2. Login to Railway

railway login

3. Link Your Local Project to Railway

In your project directory:

railway link

Follow the prompts to select your Railway project.

4. Run Commands in the Railway Environment

You can now run commands in your Railway service environment using:

railway run python main.py memory
railway run python main.py search <query>
railway run python main.py reset

This will use your Railway environment variables and database, just like in production.

---

Deployment

• Railway/Render: Deploy main.py as a worker and app.py as a web service
• Add environment variables to both services
• (Optional) Use persistent volumes for long-term memory retention
• See DEPLOYMENT_GUIDE.md for details

---

Memory System

• All prompts and responses are stored in a local SQLite database (notion_bot_memory.db)
• The bot uses the last N prompt/response pairs as context for ChatGPT (configurable via CONTEXT_WINDOW)
• Automatic memory reset: If no prompt is processed for a configurable period (INACTIVITY_RESET_HOURS), the memory is reset automatically.
• For a full technical explanation, see MEMORY_SYSTEM.md

---

File Structure

notion-ai-bot/
├── main.py              # Worker: Notion processor
├── app.py               # Web dashboard
├── memory_db.py         # SQLite memory system
├── extract_code.py      # Code extraction utility
├── requirements.txt     # Python dependencies
├── NOTION_SETUP.md      # Notion setup guide
├── DEPLOYMENT_GUIDE.md  # Deployment instructions
├── MEMORY_SYSTEM.md     # Memory system details
└── ...

---

© 2025 Germaine Luah