Skip to content

Home

Jump to bottom
Jean-Baptiste THERY edited this page Jul 13, 2026 · 1 revision

Ragmir

Ragmir is an open-source, local-first retrieval toolkit for the documents and code your agents and automations need. It indexes the files you choose, stores the index inside the project, and returns cited passages through a CLI, TypeScript API, or local MCP server.

Ragmir Core is model-agnostic and does not call a model. Use the AI or automation you already have, or pair retrieval with a local consumer when no passage may leave the workstation.

Start in five commands

Ragmir requires Node.js 20 or later.

pnpm add -D @jcode.labs/ragmir
pnpm exec rgr setup
pnpm exec rgr sources add "docs/**/*.md"
pnpm exec rgr ingest
pnpm exec rgr search "Which decision changed the rollout?"

rgr setup creates ignored local state under .ragmir/. Search results include source paths, excerpts, chunk references, line ranges, and PDF pages when available.

Choose an interface

Interface Use it for
rgr CLI Setup, ingest, cited search, audit, and maintenance
TypeScript API Embed typed retrieval in a Node.js application
Local MCP server Give agents and workflows bounded project evidence
Ragmir Chat Generate cited answers with an optional local GGUF model
Ragmir TTS Render local WAV audio or explicit online MP3

How retrieval works

flowchart LR
    A["Selected project files"] --> B["Extract and redact"]
    B --> C["Chunk and index locally"]
    C --> D["Cited passages"]
    D --> E["CLI"]
    D --> F["TypeScript API"]
    D --> G["MCP agents and automations"]
    D -. optional .-> H["Local GGUF chat"]
Loading

The default local-hash retrieval path needs no account, hosted document store, or model download. Semantic embeddings, OCR, Chat, and TTS models are explicit optional setup steps.

Guides

The repository documentation remains canonical. See the README, release history, and runnable examples for the current code-backed source.

Ragmir

Project

Clone this wiki locally