Open-source agent memory layer hits 86.67% on LongMemEval without embeddings
GabrielMartinMoran · reddit · 2026-10-28
Developer GabrielMartinMoran released Mind v1.7.0, an open-source persistent memory layer for AI agents built on SQLite + FTS5 with a unified retrieval pipeline combining phrase, token, substring, fuzzy matching and relevance signals — embeddings are optional. On a stratified 120-question subset of OmniMemEval LongMemEval-S, the best run achieved 86.67% answer accuracy with RAG disabled, near top published results (not directly comparable due to sample/judging differences). The author argues agent memory depends on structured facts, decisions, temporal context, maintenance and retrieval behavior, not just semantic similarity. MIT-licensed, local-first, MCP integration. He asks: when do embeddings genuinely beat well-designed lexical retrieval?
Related event: Open-Source Memory Layer Mind Hits 86.67% Without Vector Embeddings(2 posts)→
More from coding & agent
- Managing multiple coding agents: hourly-granularity tasks and e2e-only acceptance — dotey · 2026-10-28
- Supercharging LLM classification with logprobs: soft classifiers you can threshold-tune — JnBrymn · 2026-10-28
- Codex daily build series day 23: skipping iCloud plugins for a reset — dotey · 2026-10-28
- Cloudflare Access adds fully customizable pages with HTML + Liquid templates — dinasaur_404 · 2026-10-28
- Opus retried every session 20.6 million times during a memory leak hunt — johnlindquist · 2026-10-28
- Open-Source Skill Teaches AI Agents to Design and Produce Videos — songguoxiansen · 2026-10-28