Flaex Logo
Flaex.ai
Back to MCP Directory
Ragdocs
Model Context Protocol moderate risk

Ragdocs

RAG server providing documentation search and retrieval via vector embeddings, enabling context-aware AI responses using OpenAI and Qdrant.

Connections & Capabilities

Connects To

GitHub

Capabilities

readwrite

Quickstart

Config

{
  "mcpServers": {
    "rag-docs": {
      "command": "npx",
      "args": [
        "-y",
        "@hannesrudolph/mcp-ragdocs"
      ],
      "env": {
        "OPENAI_API_KEY": "",
        "QDRANT_URL": "",
        "QDRANT_API_KEY": ""
      }
    }
  }
}

Exposed MCP Tools (7)

safe
search_documentation

Searches the stored documentation and returns relevant excerpts based on a text query.

Read-only operation, no data modification.

safe
list_sources

Lists all documentation sources currently stored in the system.

Read-only operation, no data modification.

moderate
extract_urls

Extracts URLs from a given webpage and optionally adds them to the processing queue.

Adds URLs to a queue for later processing, potential for unintended data ingestion.

high
remove_documentation

Removes specific documentation sources from the system by their URLs.

Deletes data, impacting future search results.

safe
list_queue

Lists all URLs currently waiting in the documentation processing queue.

Read-only operation, no data modification.

moderate
run_queue

Processes and indexes all URLs currently in the documentation queue.

Processes URLs, potentially ingesting and indexing data.

moderate
clear_queue

Removes all pending URLs from the documentation processing queue.

Removes URLs from the queue, preventing them from being processed.

Safety Assessment

This server provides both read and write capabilities related to documentation. While API keys are used for authentication, the ability to add and remove documentation sources, combined with the reliance on external services like OpenAI and Qdrant, introduces moderate risk. It is safe to use for information retrieval, but caution should be exercised when adding or removing documentation sources, and API keys should be securely managed.

  • Uses API keys for authentication
  • Limited write operations through queue management
  • No direct filesystem access
  • Clear separation of concerns (search vs. queue management)
  • Uses Qdrant, a dedicated vector database
  • Requires OpenAI API key, potential for misuse if compromised
  • Qdrant API key also required, risk of data exposure if compromised
  • Can add URLs to a processing queue, potentially leading to unintended data ingestion
  • Can remove documentation sources, impacting search results
  • No explicit RBAC or access controls beyond API keys