This web application allows users to connect to their personal chat accounts (currently supporting Telegram, Webex, and Reddit) to generate AI-powered summaries or ask specific questions about their chat history within a selected date range.
The application features a robust caching system to ensure fast, repeated analysis and a flexible AI backend that can connect to multiple LLM providers.
- Multi-Service Connectivity: Securely log in to Telegram, Webex, and Reddit accounts.
- Multi-Session Support: Stay logged into all services simultaneously and seamlessly switch between them without needing to re-authenticate each time.
- Unified Chat Experience:
- Engage in a continuous conversation with the AI about your chat data.
- Ask follow-up questions without re-submitting the initial query.
- The AI maintains context throughout the conversation.
- Stop Generation: Cancel an in-progress AI response at any time.
- AI-Powered Analysis & Bot Integration:
- Threaded Conversation Support (Webex, Telegram, Reddit): Preserves native threading context end-to-end.
- Webex: Groups by parent/threadId.
- Telegram: Reconstructs reply-quote chains to deterministically resolve a thread root.
- Reddit: Correctly represents deeply nested comment trees with proper indentation. It supports two streamlined workflows (deep analysis of a post via subreddit selection, or a quick summary from a URL) controlled via a simple radio button interface.
- Configurable Image Analysis (All Providers): Globally enable/disable image processing from the UI and set maximum file sizes.
- Webex & Telegram Bot Support: Register bots to invoke the analyzer directly from any chat space.
- Real-time Streaming: View AI responses as they are generated, word-by-word.
- Summarization & Q&A: Generate concise summaries or ask specific questions about the chat content.
- Threaded Conversation Support (Webex, Telegram, Reddit): Preserves native threading context end-to-end.
- Flexible AI Model Support:
- Connects to Google AI, LM Studio, and any OpenAI-compatible endpoint.
- Dynamically populates and allows selection from all available models.
- Intelligent Caching System:
- Dramatically speeds up analysis of historical data.
- Always fetches fresh data for "today" to ensure summaries are up-to-date.
- Performance Optimizations:
- Parallel Image Downloads: Multiple images download concurrently for faster analysis (Webex & Telegram).
- Parallel Date Range Fetching: Large date ranges are split into chunks and fetched in parallel (up to 3x faster).
- Webex: Uses token-based auth for multi-client parallelization.
- Telegram: Uses shared client approach to avoid SQLite locking while enabling parallel chunk fetching.
- Configurable Chunking: Tune
parallel_fetch_chunk_days(default: 7) andmax_concurrent_fetches(default: 5) in config.json for optimal performance.
- User-Friendly Interface:
- Searchable chat list, modern date picker, and global image options.
- Flexible Downloads: Export results in multiple formats:
- Text (.txt): text-only.
- PDF (.pdf): includes embedded images.
- HTML (.html): includes images inline via data URIs.
- ZIP (.zip): bundle with transcript.txt, transcript_with_images.html (references images/), images/ files, and manifest.json metadata.
- Bot Management UI: A simple interface to register, view, and delete your bots.
- Automated Webhook Setup: Simplifies bot setup by automatically registering webhooks.
For detailed information on how to set up, configure, and use the application, please see the guides in the docs/ folder:
- Installation Guide: Detailed instructions for setting up the application using Docker or Python.
- User Guide: A complete walkthrough of the application's features.
- Webex Bot Guide: How to create and use the Webex bot.
- Telegram Bot Guide: How to create and use the Telegram bot.
- Reddit Guide: How to connect and use the Reddit integration.
- Reddit Sorting & Filtering: Advanced configuration for Reddit post sorting and filtering.
- Reddit Favorites: Using Reddit favorites for quick subreddit access.
- Technical Overview: A high-level look at the project's architecture.
A command-line interface is available to manage bot registrations.
- Add a bot:
python bot_cli.py add <backend> <name> <user_id>
- List bots:
python bot_cli.py list <user_id> [--backend <backend>]
- Remove a bot:
python bot_cli.py remove <user_id> <backend> <name>
/
├── docs/ # All documentation files
├── cache/ # Auto-generated: Stores cached messages
├── sessions/ # Auto-generated: Stores login session files
├── clients/ # Logic for communicating with chat services
├── llm/ # Logic for communicating with LLM providers
├── routers/ # API endpoint definitions (FastAPI)
├── services/ # Core business logic
├── static/ # Frontend HTML, CSS, and JavaScript
├── app.py # Main FastAPI application
├── bot_manager.py # Handles bot configuration persistence
├── config.json # User-created: Main configuration file
├── requirements.txt # Python dependencies
├── Dockerfile # Instructions to build the Docker image
└── docker-compose.yaml # Service definition for pre-built Docker Hub image