Claude Qdrant

v1.0.0Databasesstable

Local-first TypeScript MCP server for Qdrant with client isolation, LM Studio integration, and scalable document workflows.

claudeclaude-compliantclaude-desktoplm-studiolmstudio
Share:
13
Stars
0
Downloads
0
Weekly
0/5

What is Claude Qdrant?

Claude Qdrant is a Model Context Protocol (MCP) server that allows AI assistants like Claude, Cursor, and VS Code to local-first typescript mcp server for qdrant with client isolation, lm studio integration, and scalable document workflows.

Local-first TypeScript MCP server for Qdrant with client isolation, LM Studio integration, and scalable document workflows.

This server falls under the Databases category on MCPgee, the world's largest MCP server directory with 33,000+ servers.

Features

  • Local-first TypeScript MCP server for Qdrant with client iso

Use Cases

Manage vector embeddings in Qdrant with LM Studio integration.
Scale document workflows and semantic search with local-first architecture.
marlian

Maintainer

LicenseMIT
Languagejavascript
Versionv1.0.0
UpdatedApr 25, 2026
Statushealthy
Maintenanceactive

Works with

ClaudeOpenAIwindowsmacoslinux

Installation

Manual Installation

npx claude-qdrant

Configuration

Configuration Details

Config File

claude_desktop_config.json

Performance

Response Metrics

Response Time< 200ms
ThroughputMedium

Resource Usage

Memory UsageLow
CPU UsageLow

How to Set Up and Use Claude Qdrant

Claude Qdrant MCP is a local-first TypeScript MCP server that bridges Claude and other AI clients with a Qdrant vector database, enabling semantic search and RAG (retrieval-augmented generation) workflows entirely on your own hardware. It integrates with LM Studio to generate embeddings using models like BGE-M3 and supports per-client collection isolation, making it ideal for managing separate knowledge bases for different projects or customers. Developers use it to build private, scalable document retrieval systems without sending data to third-party embedding APIs.

Prerequisites

  • Node.js 18+ installed
  • Qdrant server running locally (Docker recommended) or a Qdrant Cloud instance
  • LM Studio installed and running with a BGE-M3 embedding model and an LLM such as Qwen3-8B loaded
  • npm available in your PATH
  • An MCP-compatible client such as Claude Desktop
1

Install the claude-qdrant-mcp package

Install the package globally so the `qdrant-setup` wizard and MCP server command are available.

npm install -g claude-qdrant-mcp
2

Create a project directory and run setup

Create a dedicated directory for your RAG project and run the interactive setup wizard which generates a .env file with all required configuration.

mkdir my-rag-project && cd my-rag-project
qdrant-setup
3

Configure the .env file

Edit the generated .env file with your Qdrant URL, LM Studio endpoints, embedding model, and client collection names.

QDRANT_URL=http://localhost:6333
QDRANT_API_KEY=your-api-key-if-using-cloud
LM_STUDIO_URL=http://127.0.0.1:1235
EMBEDDING_MODEL=text-embedding-finetuned-bge-m3
EMBEDDING_DIM=1024
LLM_MODEL=qwen/qwen3-8b
CLIENT_COLLECTIONS=client_a,client_b,personal,work
CONCURRENCY=5
BATCH_SIZE=10
CHUNK_SIZE=500
CHUNK_OVERLAP=10
4

Seed documents into a collection

Use the seed command to ingest documents from a local directory into a named client collection. The server chunks, embeds, and stores them in Qdrant automatically.

npm run seed -- --client work --filesdir /path/to/your/documents
# Overwrite existing collection:
npm run seed -- --client personal --filesdir /docs --overwrite
5

Test the connection

Verify Qdrant and LM Studio are reachable and collections are populated before connecting your MCP client.

npm run test-connection
6

Add to your MCP client configuration

Register the MCP server in Claude Desktop so it launches when you start a session.

Claude Qdrant Examples

Client configuration

Claude Desktop configuration for the claude-qdrant MCP server pointing at your project directory.

{
  "mcpServers": {
    "claude-qdrant": {
      "command": "npx",
      "args": ["claude-qdrant-mcp"],
      "env": {
        "QDRANT_URL": "http://localhost:6333",
        "LM_STUDIO_URL": "http://127.0.0.1:1235",
        "EMBEDDING_MODEL": "text-embedding-finetuned-bge-m3",
        "EMBEDDING_DIM": "1024",
        "CLIENT_COLLECTIONS": "work,personal,research"
      }
    }
  }
}

Prompts to try

Example prompts that leverage the semantic search and retrieval tools exposed by this server.

- "Search the 'work' collection for documents about quarterly budget planning."
- "Find all document chunks mentioning API rate limits across all my collections."
- "What does the 'research' collection contain? Show me collection statistics."
- "Search the 'personal' collection for notes about my trip to Japan."
- "Find document summaries related to machine learning across all clients."

Troubleshooting Claude Qdrant

Connection to Qdrant fails when starting the server

Ensure Qdrant is running: `docker run -p 6333:6333 qdrant/qdrant`. Confirm QDRANT_URL in .env matches the port you started Qdrant on. For Qdrant Cloud, verify QDRANT_API_KEY is set correctly.

Embedding generation fails or returns errors

Open LM Studio and confirm the BGE-M3 embedding model is loaded and the local server is active on the port specified in LM_STUDIO_URL. The embedding endpoint must be running before seeding or querying documents.

Searches return no results after seeding documents

Run `npm run seed -- --validate-only --client <name> --filesdir ./docs` to check that documents were ingested. Confirm EMBEDDING_DIM matches the dimension the BGE-M3 model actually outputs (default 1024).

Frequently Asked Questions about Claude Qdrant

What is Claude Qdrant?

Claude Qdrant is a Model Context Protocol (MCP) server that local-first typescript mcp server for qdrant with client isolation, lm studio integration, and scalable document workflows. It connects AI assistants to external tools and data sources through a standardized interface.

How do I install Claude Qdrant?

Follow the installation instructions on the Claude Qdrant GitHub repository. Clone the repo, install dependencies, and add the server config to your AI client.

Which AI clients work with Claude Qdrant?

Claude Qdrant works with all major MCP-compatible AI clients including Claude Desktop, Claude Code, Cursor, VS Code (GitHub Copilot), Windsurf, and Cline.

Is Claude Qdrant free to use?

Yes, Claude Qdrant is open source and available under the MIT license. You can use it freely in both personal and commercial projects.

Browse More Databases MCP Servers

Explore all databases servers available in the MCPgee directory. Each server includes setup guides for Claude, Cursor, and VS Code.

Quick Config Preview

{ "mcpServers": { "claude-qdrant": { "command": "npx", "args": ["-y", "claude-qdrant"] } } }

Add this to your claude_desktop_config.json or .cursor/mcp.json

Read the full setup guide →

Ready to use Claude Qdrant?

Browse our complete directory of 33,000+ MCP servers, read setup guides for your editor, and start building with the Model Context Protocol.

33,000+ ServersFree & Open SourceStep-by-Step Guides