corpus

monahand1023/corpus
★ 0 stars Python 🤖 AI/LLM Updated 1d ago
Ask natural-language questions over your own notes, PDFs, and docs. Hybrid semantic+keyword search, auto-tuned fusion, multi-hop reference expansion, and a 7-tool MCP server for Claude Code. Runs entirely locally on SQLite — no Docker, no SaaS. `pip install corpus-rag`
View on GitHub → Try with Claude — $10 free →

Quick Install

Copy the config for your editor. Some servers may need additional setup — check the README.

Add to claude_desktop_config.json:

{
  "mcpServers": {
    "corpus": {
      "command": "uvx",
      "args": [
        "corpus"
      ]
    }
  }
}

Or install with pip: pip install corpus

README Excerpt

Your personal archive — notes, PDFs, docs — queryable in plain English, running entirely on your machine. A personal knowledge system shouldn't require a vector database service, a SaaS subscription, or trusting your files to someone else's cloud. `corpus` is one Python process, one SQLite file, one MCP server. Add a `corpus.toml`, point it at your data, run `corpus-ingest`, and Claude Code can search years of notes in under 300ms.

Tools (13)

corpus_statsexpand_contextget_docget_summaryhtmlmarkdownnotespdfrecent_activitysearch_knowledgetexttimelinetype

Topics

claude-codeembeddingsgeminimcppythonragsqlite-vecvoyage-ai