#2727th of 27 in Memory
memory-lancedb-pro
LanceDB-backed long-term memory plugin for OpenClaw agents
- Stars
- 4.5k
- License
- no license
- Last commit
- Nov 2026
- Last release
- Mar 2026
- Language
- JavaScript
4,092 on 22 Apr 20264,450 on 11 Nov 2026
From GitHub's daily star counts.
Checked on GitHub.
Overview
OpenClaw plugin that stores preferences, decisions and project context in LanceDB and injects relevant memories before each reply. It auto-captures with LLM-based extraction into 6 categories, retrieves with vector plus BM25 search and cross-encoder reranking, and applies Weibull decay to fade low-value memories. Embeddings come from OpenAI, Jina, Gemini, Ollama or any OpenAI-compatible API.
Who it is for: OpenClaw users who want persistent, searchable memory across agent sessions
Strengths
- Hybrid retrieval: vector plus BM25, fused and reranked with a cross-encoder
- Per-agent, per-user and per-project scope isolation
- Works with any OpenAI-compatible embedding API, including local Ollama
- CLI covers export, import, upgrade, reembed and migration from built-in memory-lancedb
Weaknesses
- Tied to OpenClaw 2026.3+ plugin hooks; not usable with other agent hosts
- Requires AVX CPU; AVX2 builds can crash with SIGILL unless native cosine is disabled
- Changing embedding model or dimensions requires re-embedding into a fresh database
- Only 4 core tools exposed by default; full toolset needs enableManagementTools
What it needs
- no GPU
- Needs OpenClaw 2026.3+, LanceDB, Embedding provider API (OpenAI, Jina, Gemini, Ollama or compatible), AVX-capable CPU
- Models: OpenAI, Jina, Gemini, Ollama, OpenAI-compatible APIs
What changed
- v1.0.32 ↗ (opens in a new tab)Breaking
From the release notes. No flag means the note does not say so, not that the upgrade is safe.
Also in Memory
See all 27| Rank | Project | Score |
|---|---|---|
| 1 | hindsightAgent memory server with retain, recall and reflect operations | 95 out of 100 |
| 2 | gbrainMemory store with sources and a knowledge graph for coding agents | 91 out of 100 |
| 3 | MemPalaceLocal verbatim memory store for AI agents with semantic search and MCP | 90 out of 100 |
| #4 | SupermemoryMemory and context engine for AI agents, with user profiles and hybrid search | 89 out of 100 |
| #5 | Mem0Long-term memory layer that stores and retrieves user facts for LLM agents | 88 out of 100 |