Better Qdrant

Provides MCP endpoints to manage Qdrant collections, add documents with embeddings, and perform semantic searches.

typescript

GitHub Stars

typescript

Language

6 months ago

First Indexed

2 months ago

Catalog Refreshed

Documentation & install

Readme and setup notes from the catalogue, plus a client-ready config you can copy for your MCP host.

Installation

Add the following to your MCP client configuration file.

Configuration

View docs

{
  "mcpServers": {
    "wrediam-better-qdrant-mcp-server": {
      "command": "npx",
      "args": [
        "better-qdrant-mcp-server"
      ],
      "env": {
        "QDRANT_URL": "http://localhost:6333",
        "OLLAMA_MODEL": "nomic-embed-text",
        "OPENAI_API_KEY": "your_openai_api_key",
        "QDRANT_API_KEY": "your_api_key_if_needed",
        "OLLAMA_ENDPOINT": "http://localhost:11434",
        "OPENROUTER_API_KEY": "your_openrouter_api_key"
      }
    }
  }
}

You run the Better Qdrant MCP Server to manage Qdrant vector collections, add documents, and run semantic searches from a single MCP-enabled workflow. This server exposes practical tools to list collections, ingest documents via multiple embedding services, perform semantic queries, and remove collections as needed.

How to use

Use an MCP client to interact with the Better Qdrant MCP Server. You can list available collections, add documents to a specific collection using your chosen embedding service, perform semantic searches across your vectors, and delete a collection when it is no longer needed. Each operation is exposed as a distinct tool you invoke through your MCP client, making it easy to compose workflows like indexing new documents and then querying them via natural language.

How to install

Prerequisites: you need Node.js version 18.0.0 or newer. Ensure you have a running Qdrant server (local or remote) and any embedding service API keys you plan to use.

npm install -g better-qdrant-mcp-server

Or run it directly with npx:

npx better-qdrant-mcp-server

Additional notes

After you start the MCP server, you can configure it to connect to Qdrant and embedding services using environment variables. A typical setup includes QDRANT_URL, QDRANT_API_KEY (if needed), and keys for the embedding providers you plan to use. You can also specify default embedding service behavior to streamline your workflows.

Configuration

Configure the MCP server with environment variables in a .env file at your project root. The following are examples you may include:

# Qdrant Configuration
QDRANT_URL=http://localhost:6333
QDRANT_API_KEY=your_api_key_if_needed

# Embedding Service API Keys
OPENAI_API_KEY=your_openai_api_key
OPENROUTER_API_KEY=your_openrouter_api_key
OLLAMA_ENDPOINT=http://localhost:11434

Examples and usage notes

To use this MCP server with Claude, you can set up an MCP configuration that points to the server and defines embedding behavior, including which service to use by default and necessary API keys. This enables seamless indexing and querying workflows across your data.

Troubleshooting

If you encounter connection issues to Qdrant, verify that QDRANT_URL is correct and that the Qdrant server is reachable from the MCP server host. Check that embedding service keys (e.g., OPENAI_API_KEY, OPENROUTER_API_KEY) are valid and that any local embedding endpoints (such as Ollama) are running if you rely on local models.

Notes

The server supports multiple embedding services, including OpenAI, OpenRouter, Ollama, and FastEmbed. You can mix services across operations as needed, though you may choose a default embedding service to streamline typical workflows.

Tools overview

List Collections: View all available Qdrant collections
Add Documents: Process and add documents to a Qdrant collection with various embedding services
Search: Perform semantic searches across your vector database
Delete Collection: Remove collections from your Qdrant database

Available tools

list_collections

List all available Qdrant collections to understand your data structure and plan further actions.

add_documents

Process and add documents to a specific Qdrant collection using a chosen embedding service, with options for chunking and overlap.

search

Perform semantic searches across your vector database by providing a query, target collection, embedding service, and limit for results.

delete_collection

Remove a collection from your Qdrant database to clean up unused data.

Built by

VeilStrat

AI signals for GTM teams