Mimirs – persistent local memory for AI coding agents (MCP)
76% token reduction on real projects with one-command setup.
High-performance MCP memory server for multi-agent systems — SQLite-backed with hybrid search.
SQLite-backed MCP server gives local agents shared memory without vector DB overhead.
Developers building multi-agent systems using MCP
Zep · Mem0 · Supermemory
76% token reduction on real projects with one-command setup.
Local memory for AI agents that actually learns from your last 50 sessions, not just context window tricks.
Local RAG for Claude when Cursor and other editors already handle context.
Local-first design — SQLite plus on-device ONNX embeddings — provides a small-footprint way to share operational lessons between agents, and automatic PII redaction is a thoughtful, tangible safeguard. The one-line publish/query API, cross-language SDKs (Python + TypeScript), and a planned FastAPI/Postgres+pgvector server show practical engineering trade-offs rather than vaporware.
Three-tier memory architecture solves context bloat for Claude Code and Copilot CLI users.
Predict-calibrate extraction reduces noise, but Zep and Mem0 already dominate the agent memory space.