Vibe Coding Discovermira
View on GitHuba best-effort approximation of a synthetic entity complete with all of the primatives required for long-horizon persistance. batteries included.
MIRA is a self-hosted conversational agent built around persistent memory and long-running context. It uses PostgreSQL/pgvector for retrieval, automatically summarizes conversation segments, adapts its behavior from feedback, and supports dynamically loaded tools and documents.
Use Cases
Run a self-hosted persistent AI assistantRecall relevant facts across long-running conversationsSearch and manage personal memoriesCollaborate on versioned documentsUse dynamically loaded tools for tasksAdapt assistant behavior from user feedback
Built With
- Language
- Python
- Frameworks
- FastAPI · PostgreSQL · pgvector · Anthropic SDK · MCP
Tags
AI assistant · long-term memory · persistent agent · RAG · semantic search · self-hosted · tool use · context management · personalization · multi-provider