Index local Claude/Codex history and output JSONL search results.
npx skills add https://github.com/voznik/dotfiles --skill memex-search-voznik
Please help me install this Agent Skill. Skill: memex-search Source: https://github.com/voznik/dotfiles/tree/main/.rulesync/skills/memex-search Command: npx skills add https://github.com/voznik/dotfiles --skill memex-search-voznik
Memex Search indexes local Claude/Codex history and retrieves results in a structured, machine-friendly JSON format for efficient search, analysis, and downstream processing.
Index your Claude/Codex history with memex index, then execute memex search to retrieve JSON-formatted results.
High-intent search queries and answers about installing and using this skill.
You can index local Claude history for RAG by running a full or incremental reindexing process. The pipeline generates JSONL output with embeddings enabled by default, making the historical data structured and ready for retrieval.
Yes, you can filter search results by session, project, role, and source. This allows you to narrow down the structured JSON output to specific contexts before feeding it into downstream processing.
The best way to perform semantic search on Codex logs is using a search tool that supports embeddings and multiple retrieval modes. Lexical, semantic, and hybrid search are all supported natively.
Yes, incremental indexing supports embeddings by default. This ensures your searchable index stays updated with new history entries without requiring a full rebuild every time.
You get structured JSONL output by indexing your Claude history and executing a search query. The pipeline outputs JSONL formatted results suitable for downstream RAG pipelines and analysis.
You need hybrid search when simple keyword matching is insufficient and semantic context matters. It combines lexical and semantic search modes to retrieve the most relevant results from your indexed history.
kvsankarSearch Claude Code conversation history across local, WSL, Windows, and SSH remotes.
fableindigo-gifPerform semantic vector search against AgentDB stores with HNSW indexing.
bcastelinoDesign and optimize RAG systems for LLM applications.
lookleeEnable session debug logging and analyze logs for errors and warnings.
