3 comments
Everybody builds one. And, then they usually figure out that making the model fill its context with a bunch of memories hurts performance more often than it helps.
Given the abundance of vaguely similar local-first AI memory layers, it might be a good idea to add a "Why Mnemo" section right at the top of README.md to explain why folks should consider using it.
You forgot BM25 embeddings.<p><a href="https://github.com/MikeS071/ai-engram" rel="nofollow">https://github.com/MikeS071/ai-engram</a><p><a href="https://github.com/lamost423/openclaw-hybrid-memory" rel="nofollow">https://github.com/lamost423/openclaw-hybrid-memory</a><p><a href="https://medium.com/@qdrddr/agentic-memory-framework-hindsight-with-fts-bm25-hybrid-search-and-rabitq-ce8141a0323e" rel="nofollow">https://medium.com/@qdrddr/agentic-memory-framework-hindsigh...</a><p><a href="https://clawhub.ai/vnesin-sarai/hybrid-retrieval" rel="nofollow">https://clawhub.ai/vnesin-sarai/hybrid-retrieval</a><p><a href="https://www.josecasanova.com/blog/openclaw-qmd-memory" rel="nofollow">https://www.josecasanova.com/blog/openclaw-qmd-memory</a><p><a href="https://medium.com/@richardhightower/stop-the-hallucinations-hybrid-retrieval-with-bm25-pgvector-embedding-rerank-llm-rubric-rerank-895d8f7c7242" rel="nofollow">https://medium.com/@richardhightower/stop-the-hallucinations...</a><p><a href="https://github.com/oomkapwn/enquire-mcp#-why-its-the-best" rel="nofollow">https://github.com/oomkapwn/enquire-mcp#-why-its-the-best</a><p><a href="https://github.com/rohitg00/agentmemory#key-capabilities" rel="nofollow">https://github.com/rohitg00/agentmemory#key-capabilities</a><p><a href="https://github.com/Melody-0321/NE-Memory-Core" rel="nofollow">https://github.com/Melody-0321/NE-Memory-Core</a><p><a href="https://github.com/ClaudioDrews/memory-os" rel="nofollow">https://github.com/ClaudioDrews/memory-os</a><p><a href="https://en.wikipedia.org/wiki/Okapi_BM25" rel="nofollow">https://en.wikipedia.org/wiki/Okapi_BM25</a><p>> It is based on the probabilistic retrieval framework developed in the 1970s and 1980s<p>Anyway, good for ya, hope you had fun building it.