MCP Catalogs
← Home

minima

by dmayboroda·★ 1,048·Score 52

Minima is an on-premises RAG system with MCP integration for secure local document querying.

ai-llmknowledge-graphproductivity
103
Forks
14
Open issues
4 mo ago
Last commit
2d ago
Indexed

Overview

Minima is an open-source RAG system that runs on-premises in containers, with built-in MCP support. It can operate fully locally with Ollama or integrate with various LLM providers including ChatGPT, Anthropic Claude, or custom OpenAI-compatible APIs. The system indexes documents and allows users to query them through multiple interfaces, ensuring data privacy while maintaining flexibility in deployment options.

Try asking AI

After installing, here are 5 things you can ask your AI assistant:

you:Secure document retrieval for enterprises with strict data privacy requirements
you:Personal knowledge base with local document search capabilities
you:Integration with Claude Desktop via MCP for enhanced AI assistants
you:What file types does Minima support?
you:Can Minima be used with custom LLM servers?

When to choose this

Choose Minima when you need secure, on-premises document retrieval with RAG capabilities and want to integrate with AI assistants like Claude or ChatGPT.

When NOT to choose this

Don't choose Minima if you need advanced document formatting preservation, require real-time collaboration features, or need a fully managed cloud service without self-hosting requirements.

Tools this server exposes

3 tools extracted from the README (low confidence)
  • query_documents

    Search and retrieve information from local documents

  • index_files

    Index local files for retrieval-augmented generation

  • search_local_files

    Search through indexed local files for relevant information

Note: Tool names inferred from documentation as specific MCP tools aren't explicitly named. The server appears to document functionality rather than explicit tool names.

Comparable tools

private-gptanythingllmlangchain

Installation

Docker Installation

  1. Clone the repository and navigate to it
  2. Create a .env file using .env.sample as template
  3. Set required variables (LOCAL_FILES_PATH, EMBEDDING_MODEL_ID, EMBEDDING_SIZE)
  4. Run: docker compose -f docker-compose-mcp.yml --env-file .env up --build

Claude Desktop Integration

Add to your claude_desktop_config.json:

{
  "mcpServers": {
    "minima": {
      "command": "uv",
      "args": [
        "--directory",
        "/path_to_cloned_minima_project/mcp-server",
        "run",
        "minima"
      ]
    }
  }
}

GitHub Copilot Integration

Create or update .vscode/mcp.json with:

{
  "servers": {
    "minima": {
      "type": "stdio",
      "command": "path_to_cloned_minima_project/run_in_copilot.sh",
      "args": [
        "path_to_cloned_minima_project"
      ]
    }
  }
}

FAQ

What file types does Minima support?
Minima supports .pdf, .xls, .docx, .txt, .md, .csv file types for indexing and retrieval.
Can Minima be used with custom LLM servers?
Yes, Minima supports custom LLM servers through OpenAI-compatible APIs including vLLM, TGI, Ollama server, LiteLLM, LocalAI, and any OpenAI-compatible endpoint.

On Hacker News

Recent discussion from the developer community.

Compare minima with

minima vs ultimate_mcp_serverminima vs mcp-server-chartminima vs everythingminima vs filesystemminima vs time
GitHub →

Last updated · Auto-generated from public README + GitHub signals.