Mimir

Getting Started

Quick setup guide to get Mimir running in minutes

Get Mimir up and running in just a few steps.

Prerequisites

Before you begin, make sure you have:

  • Node.js: 20 or later
  • PostgreSQL Database: A PostgreSQL database with pgvector extension (Supabase, Neon, self-hosted, etc.)
  • LLM Provider: API key for OpenAI, Anthropic, Google, or Mistral
  • GitHub (optional): Repository with documentation (MDX) and/or code to ingest

Step 1: Set up PostgreSQL Database

You need a PostgreSQL database with the pgvector extension. You can use any provider:

  • Supabase – Create a project, then use Settings → Database for your connection string. With Docker, use the Session Pooler (port 6543) if your database shows "Not IPv4 compatible".
  • Neon – Create a project and copy the connection string from the dashboard.
  • Self-hosted – Install PostgreSQL and the pgvector extension, then use a connection string like postgresql://user:password@host:5432/database.

Save your connection string; you'll add it as MIMIR_DATABASE_URL in the next step.

Step 2: Configure mimir-rag

cd mimir-rag
cp .env.example .env
# Edit .env with your PostgreSQL database and LLM credentials

See the Configuration guide for all available options.

Step 3: Generate API key

Generate a secure API key for your server:

npm run generate-apikey

This will output an API key that you'll use to authenticate requests to your mimir-rag server.

Step 4: Start the server

make server

The server will:

  • Load your .env configuration
  • Run Prisma migrations to set up the database schema (if MIMIR_DATABASE_URL is set)
  • Start the API on port 3000

Step 5: Trigger ingestion

Ingest your first documentation or codebase:

npm run ingest:cli

This will fetch files from your configured GitHub repositories, extract entities, and store them in your PostgreSQL database.

Next Steps

Once you have mimir-rag running:

  1. Check out the Configuration guide to learn about all available environment variables
  2. Learn how to configure multiple repositories
  3. Set up MCP integration to connect AI assistants
  4. Explore the API Reference for programmatic access
PreviousIntroduction
NextConfiguration