aiming-lab/SimpleMem

SimpleMem: Efficient Lifelong Memory for LLM Agents

81
/ 100
Verified

Implements a three-stage semantic compression pipeline—structured compression, online synthesis, and intent-aware retrieval—to maximize information density while minimizing token overhead. Exposes memory functionality through MCP (Model Context Protocol) servers and Python packages, integrating with Claude Desktop, Cursor, LM Studio, and other AI platforms. Supports persistent cross-session memory that reportedly outperforms Claude's native memory by 64% on standard benchmarks.

3,182 stars and 2,317 monthly downloads. Used by 1 other package. Actively maintained with 11 commits in the last 30 days. Available on PyPI.

Maintenance 20 / 25
Adoption 19 / 25
Maturity 22 / 25
Community 20 / 25

How are scores calculated?

Stars

3,182

Forks

310

Language

Python

License

MIT

Last pushed

Mar 10, 2026

Monthly downloads

2,317

Commits (30d)

11

Dependencies

8

Reverse dependents

1

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/embeddings/aiming-lab/SimpleMem"

Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.