Run local hybrid searches with BM25, embeddings, and LLM re-ranking.
npx skills add https://github.com/xintuchain/tongtong --skill qmd-memory
Please help me install this Agent Skill. Skill: qmd-memory Source: https://github.com/xintuchain/tongtong/tree/main/skills/qmd-memory Command: npx skills add https://github.com/xintuchain/tongtong --skill qmd-memory
QMD Memory eliminates recurring API costs and latency caused by remote memory and embedding lookups by running hybrid search locally and serving semantic context to OpenClaw agents.
Run the OpenClaw skill setup to install QMD, configure collections, and generate local embeddings for your workspace.
High-intent search queries and answers about installing and using this skill.
You can eliminate recurring API memory search costs by performing local hybrid searches using QMD, which combines BM25 SQLite FTS5 indexing, local vector embeddings, and LLM re-ranking to retrieve context without external API calls.
Local hybrid search works by combining BM25 keyword matching through SQLite FTS5, local vector embeddings for semantic similarity, and LLM re-ranking to return high-quality context from indexed collections.
Yes, a Node.js environment is required to generate local embeddings and run the QMD memory indexing process for your workspace, daily logs, and project collections.
Yes, you can share a single memory index across multiple agents by enabling the optional MCP server, which supports multi-agent sharing and collaborative querying of past decisions and research.
The best way to index workspace files, daily logs, intelligence, and project documents is using the nightly indexing feature, which auto-configures collections and generates local embeddings for precise retrieval.
Remote memory lookup causes latency through external network requests and embedding API charges, which you can avoid by running local hybrid search and serving semantic context directly to OpenClaw agents.
zircoteConfigure the qmd CLI for semantic search indexing and embedding generation.
steipeteIndex local files and search them with BM25, vector, and hybrid queries.
SafphereCreate new Claude Skills with structured SKILL.md frontmatter and optional resources.
profbernardojDiagnose and repair OpenClaw memory search with local embeddings and hybrid search.
