RagGo Documentation

Welcome to the RagGo docs. RagGo is a self-hosted, privacy-first RAG system built as a Go MCP server with a Python gRPC microservice. It runs entirely on your own hardware.

New here? Start with the Quickstart guide to get RagGo running in under 10 minutes.

What is RagGo?

RagGo gives you AI-powered semantic search over your own documents — without sending anything to the cloud. It bundles everything you need: a local embedding model (intfloat/multilingual-e5-small), a Qdrant vector database, a Web UI, and an MCP server with 16 tools.

Download, extract, and run ./raggo. It auto-starts the Python service, Qdrant, and the Web UI. Upload documents through the Web UI or connect an MCP client to search your knowledge.

Core concepts

Ingestion: Documents are chunked (up to 1000 tokens), embedded into vectors, and stored in Qdrant. Upload via the Web UI at /ui/ingest or use the ingest_file and ingest_documents_directory MCP tools.
Search: Your query is embedded and compared against stored vectors. Search via the Web UI at /ui/search or use the search_documents and search_code MCP tools.
GraphRAG: For codebases, RagGo provides a NetworkX-based knowledge graph of class/function relationships that can be queried via MCP tools.
MCP: RagGo exposes 16 MCP tools at http://127.0.0.1:3235/mcp, letting Claude, Cursor, and other AI assistants interact with your data natively.

Sections

Quickstart

Get RagGo running in 3 steps.

Installation

Platform-specific install guide.

Document Ingestion

Supported formats, chunking, Web UI & MCP ingestion.

Web UI search, MCP tools, result format.

MCP Integration

Connect Claude, Cursor, and other clients.

GraphRAG

Code knowledge graph queries.

License Activation

Activate and manage your license.

Privacy guarantee

In default mode, RagGo binds to 127.0.0.1. The only external network request is an optional one-time download of the Qdrant binary during first-run setup. After that, no data leaves your machine — embedding, storage, and search are all local. Paid tiers include AES-256-GCM data-at-rest encryption for an additional layer of security.