MCP Servers · qdrant

mcp-server-qdrant

mcp-server-qdrant is an official Model Context Protocol server that integrates Qdrant vector database with LLM applications, enabling semantic memory storage and retrieval. It provides two main tools—qdrant-store and qdrant-find—allowing LLMs to persistently store and query information as embeddings.

Talk to DEV.co Discuss Implementation

Source: GitHub — github.com/qdrant/mcp-server-qdrant

1.5k

GitHub stars

277

Forks

Python

Primary language

Apache-2.0

License (OSI-approved)

Key facts

Objective fields from the source. Values we can't verify are shown as “Unknown” rather than guessed.

Field	Value
Repository	qdrant/mcp-server-qdrant
Owner	qdrant
Primary language	Python
License	Apache-2.0 — OSI-approved
Stars	1.5k
Forks	277
Open issues	65
Latest release	v0.8.1 (2025-12-10)
Last updated	2026-06-26
Source	https://github.com/qdrant/mcp-server-qdrant

What mcp-server-qdrant is

A Python-based MCP server built on FastMCP that wraps Qdrant's vector search capabilities, supporting both remote (HTTP) and local database connections. It uses FastEmbed for text embeddings and provides configurable transport protocols (stdio, SSE, streamable-HTTP) for LLM client integration.

Quickstart

Get the mcp-server-qdrant source

Clone the repository and explore it locally.

terminalbash

1git clone https://github.com/qdrant/mcp-server-qdrant.git2cd mcp-server-qdrant3# follow the project's README for install & configuration

Need it deployed, integrated, or customized instead? DEV.co ships production installs.

Best use cases

AI Agent Semantic Memory

Enable Claude, Cursor, or custom LLM agents to store and retrieve conversation context, user preferences, or domain knowledge as persistent semantic memories across sessions.

RAG-Augmented Chat Applications

Integrate with chat interfaces to provide LLMs with relevant retrieved context from a vector database, improving response quality with grounded, searchable information.

Knowledge Base Integration for IDE Assistants

Embed into AI-powered IDEs (Cursor, Windsurf) to give code assistants access to project-specific documentation, API references, or best practices as queryable vector memories.

Implementation considerations

Embedding model selection (currently FastEmbed only) directly impacts memory footprint and query quality; sentence-transformers/all-MiniLM-L6-v2 is the default but may not suit specialized domains.
Environment variables control all configuration (Qdrant URL/local path, API key, collection name, embedding model); no QDRANT_URL and QDRANT_LOCAL_PATH cannot coexist.
Read-only mode (QDRANT_READ_ONLY) disables the store tool; plan for separate deployments if write and read operations need isolation.
Default search limit is 10 results; tune QDRANT_SEARCH_LIMIT based on context window and relevance requirements.
FastMCP dependencies (FASTMCP_SERVER_DEPENDENCIES) can be injected, enabling custom embedding or post-processing steps.

When to avoid it — and what to weigh

Keyword-Based Search Required — If your use case demands traditional full-text or keyword search rather than semantic similarity, a vector database adds unnecessary complexity.
Sub-Millisecond Latency Critical — For ultra-low-latency applications, the overhead of embedding generation and Qdrant queries may introduce unacceptable delays depending on model and infrastructure.
Structured Data Queries Only — If you need complex relational queries, ACID transactions, or SQL-style joins, a traditional SQL database paired with vector search would be more appropriate.
No Vector Database Infrastructure Available — Requires either a running Qdrant server or local disk space; if neither is feasible, this integration is not deployable.

License & commercial use

Licensed under Apache License 2.0, a permissive OSI-approved license allowing commercial and proprietary use, modification, and distribution provided the Apache header and license notice are retained.

Apache 2.0 is a permissive license suitable for commercial products. However, ensure compliance with dependency licenses (Qdrant client, FastMCP, sentence-transformers, FastEmbed) in your deployment. No warranty or liability limitations are stated in the MCP server itself; review Qdrant server licensing separately if using commercial Qdrant Cloud.

DEV.co evaluation signals

Editorial assessment — not user reviews. Directional, with an explicit confidence level.

Signal	Assessment
Maintenance	Active
Documentation	Adequate
License clarity	Clear
Deployment complexity	Low
DEV.co fit	Strong
Assessment confidence	High

Security considerations

API keys passed via environment variables; ensure secure management in CI/CD and container orchestration (Kubernetes secrets, AWS Secrets Manager, etc.). No encryption at rest documented; depends on Qdrant server configuration. Input validation for tool parameters (information, query, collection_name) is not explicitly described; assume standard sanitization. No explicit authentication between MCP client and server; rely on transport security (TLS for SSE/HTTP) and host-level access controls.

Alternatives to consider

Pinecone MCP Server

Similar semantic memory integration for LLMs but managed cloud-only (no self-hosted option); different pricing model and ecosystem.

Weaviate MCP or LangChain Vector Store

Alternative vector databases with MCP bindings or LangChain integration; more feature-rich but higher operational overhead.

Custom Flask/FastAPI Vector API

Full control and minimal overhead but requires building and maintaining MCP protocol support and embedding pipeline yourself.

Software development agency

Build on mcp-server-qdrant with DEV.co software developers

Deploy mcp-server-qdrant in minutes using uvx or Docker to give your AI agents persistent, queryable memory across sessions.

Talk to DEV.co

Related open-source tools

Surfaced by semantic similarity across the DEV.co open-source index.

mysql_mcp_server arxiv-mcp-server qdrant-client

Related on DEV.co

Explore the category and the services that help you build with it.

MCP Servers AI Development API Development Cloud Deployment Talk to DEV.co

mcp-server-qdrant FAQ

Can I use mcp-server-qdrant without a Qdrant server?

Yes, set QDRANT_LOCAL_PATH to a local directory for embedded Qdrant mode. However, QDRANT_LOCAL_PATH and QDRANT_URL cannot be set simultaneously.

What embedding models are supported?

Currently only FastEmbed models via the EMBEDDING_PROVIDER. Default is sentence-transformers/all-MiniLM-L6-v2. Support for other providers is not implemented.

How do I deploy this to production?

Use Docker with SSE or streamable-HTTP transport, set FASTMCP_SERVER_HOST='0.0.0.0', and ensure Qdrant server is externally reachable. Use environment variable injection (Kubernetes, Docker Compose secrets) for API keys and credentials.

Is this suitable for high-concurrency workloads?

Unknown; no performance benchmarks or concurrency limits are documented. Depends on underlying Qdrant server capacity and embedding model latency.

Custom software development services

Adopting mcp-server-qdrant is usually one piece of a larger software development effort. As a software development agency, DEV.co provides software development services and web development expertise — pairing senior software developers and web developers with your team to design, build, and operate mcp servers software in production.

Ready to Add Semantic Memory to Your LLM?

Deploy mcp-server-qdrant in minutes using uvx or Docker to give your AI agents persistent, queryable memory across sessions.

Talk to DEV.co Schedule a Consultation