qualixar/superlocalmemory

World's first local-only AI memory to break 74% retrieval and 60% zero-LLM on LoCoMo. No cloud, no APIs, no data leaves your machine. Additionally, mode C (LLM/Cloud) - 87.7% LoCoMo. Research-backed. arXiv: 2603.14588

55
/ 100
Established

Implements a hybrid retrieval architecture combining Fisher-Rao geodesic distance (information geometry), BM25 keyword matching, entity graph spreading activation, and temporal indexing with RRF fusion and cross-encoder reranking — entirely replacing cloud LLM dependency with mathematical guarantees from differential geometry and algebraic topology. Offers dual MCP + CLI interfaces for IDE integration and agent-native JSON output, with three operating modes (A: zero-cloud, B: local Ollama, C: optional cloud) supporting frameworks like Claude, Cursor, and custom agent pipelines. Designed for EU AI Act compliance and achieves 85% on open-domain retrieval without any cloud calls or GPU requirements.

32 stars and 4,841 monthly downloads. Used by 1 other package. Available on PyPI.

Maintenance 13 / 25
Adoption 16 / 25
Maturity 20 / 25
Community 6 / 25

How are scores calculated?

Stars

32

Forks

2

Language

Python

License

Last pushed

Mar 18, 2026

Monthly downloads

4,841

Commits (30d)

0

Dependencies

9

Reverse dependents

1

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/rag/qualixar/superlocalmemory"

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