mcp-server-qdrant

What it is and what it does

mcp-server-qdrant bridges LLM applications and Qdrant vector search by implementing the Model Context Protocol (MCP) standard. It runs as a standalone server that exposes two core tools: qdrant-store (to persist text and metadata into Qdrant collections) and qdrant-find (to retrieve semantically similar results via vector search). The server uses fastembed for embedding generation and can connect to either a remote Qdrant instance or a local database.

Configuration is entirely environment-variable driven, supporting multiple transport protocols (stdio, SSE, streamable-HTTP) and deployment modes including Docker and uvx. It integrates with Claude Desktop and other MCP clients, enabling AI applications to maintain and query a persistent semantic memory layer without direct database access.

Use it for:

• Integrate Claude Desktop with Qdrant to give AI assistants persistent semantic memory across conversations
• Build AI-powered applications that store and retrieve domain-specific context from vector embeddings
• Deploy a containerized MCP server to provide remote LLM clients with vector search capabilities
• Create custom AI workflows where multiple agents share a centralized semantic knowledge base in Qdrant

Worth the install?

AI-flagged interpretation of the facts on this page — verify before relying

Exposes Qdrant vector search as a Model Context Protocol server, allowing LLM applications to store and retrieve semantic memories from a vector database via standardized MCP tools.

Yes, if you need MCP-compatible vector search integration. The package is straightforward to deploy, has low install friction, carries a permissive license, and has no known vulnerabilities. The aging maintenance status (247 days since last release) is a minor caution but not a blocker for stable, read-write workloads. Verify active support if you plan to rely on ongoing updates.

Install

mcp-server-qdrant on PyPI

pip install mcp-server-qdrant

uv add mcp-server-qdrant

poetry add mcp-server-qdrant

Installing mcp-server-qdrant

Before you install

Low install friction with a pure-Python wheel and four runtime dependencies. Maintenance status is aging—last release was 247 days ago—but the package remains functional and supports current Python versions.

License in practice

Apache-2.0 permissive license allows commercial and private use with minimal restrictions, making it suitable for most deployment contexts.

Quickstart

# Install
pip install mcp-server-qdrant

# Set environment variables
export QDRANT_URL="http://localhost:6333"
export COLLECTION_NAME="my-collection"
export EMBEDDING_MODEL="sentence-transformers/all-MiniLM-L6-v2"

# Run the server
from mcp_server_qdrant import app
# Server exposes qdrant-store and qdrant-find tools via MCP

Requires a running Qdrant instance (remote via QDRANT_URL or local via QDRANT_LOCAL_PATH) and Python >=3.10. fastembed will download its embedding model on first use.

Verify before relying

• Whether the package is actively maintained or in maintenance-only mode beyond the aging signal
• Performance characteristics when handling large vector collections or high query volume
• Compatibility with specific Qdrant server versions or deployment modes

Package facts

Evidence: mcp_server_qdrant-0.8.1-py3-none-any.whl

Tags

More Database packages

Further reading