Bluesky Bot#

A virtual person for Bluesky powered by LLMs, built with FastAPI and pydantic-ai.

Setup#

1. Install dependencies:

just sync

2. Copy .env.example to .env and add your credentials:

• BLUESKY_HANDLE: Your bot's Bluesky handle
• BLUESKY_PASSWORD: App password from Bluesky settings
• ANTHROPIC_API_KEY: For AI responses (optional, falls back to placeholder)

3. Test posting:

just test-post

4. Run the bot:

just dev

Current Features#

• ✅ Responds to mentions with placeholder or AI messages
• ✅ Proper notification handling (no duplicates)
• ✅ Graceful shutdown for hot-reload
• ✅ AI integration with Anthropic Claude (when API key provided)
• ✅ Status page at /status showing activity and health
• 🚧 Memory system (coming soon)
• 🚧 Self-modification capabilities (planned)

Architecture#

• FastAPI web framework with async support
• pydantic-ai for LLM agent management
• TurboPuffer for scalable vector memory (planned)
• AT Protocol for Bluesky integration

Development#

just           # Show available commands
just dev       # Run with hot-reload
just test-post # Test posting capabilities
just fmt       # Format code
just status    # Check project status

Status Page#

Visit http://localhost:8000/status while the bot is running to see:

• Current bot status and uptime
• Mentions received and responses sent
• AI mode (enabled/placeholder)
• Last activity timestamps
• Error count

Memory Architecture#

See sandbox/memory_architecture_plan.md for the planned memory system using TurboPuffer.

Reference Projects#

This bot is inspired by:

• Void by Cameron Pfiffer - Sophisticated memory system
• Penelope by Hailey - Self-modifying capabilities
• Marvin Slackbot - Multi-agent architecture

See sandbox/ for detailed analysis of each project.