Persistent AI Memory

What it does

Persistent AI Memory provides a robust long-term memory substrate for AI assistants. It uses SQLite for structured storage and vector embeddings for semantic retrieval, allowing LLMs to remember user preferences, past conversations, and complex project contexts across different sessions.

Tools

• store_memory: Persistently saves important information with user and model isolation.
• search_memories: Retrieves relevant memories using semantic vector search.
• store_conversation: Captures and archives chat history for later recall.
• search_conversations: Searches through historical conversations to find specific context.
• log_tool_call: Records how the AI uses tools to enable self-reflection and pattern analysis.
• get_system_health: Validates database connectivity and embedding provider status.

Installation

Install the package via pip:

pip install git+https://github.com/savantskie/persistent-ai-memory.git

To use as an MCP server, run the server module:

python -m ai_memory_mcp_server

Supported hosts

• claude
• openwebui