Qdrant MCP
by Qdrant
Semantic memory layer for AI agents with vector search via Qdrant
database Python Beginner Self-hostable No API key Docker Verified
⭐ 1.2k stars 📅 Updated: 1mo ago
Description
Official Qdrant MCP server that functions as a semantic memory layer for AI agents. Provides a deliberately simple two-tool interface for storing information with metadata and performing semantic searches across stored knowledge. Designed to give LLMs persistent memory that survives across conversations by persisting data in Qdrant's vector database. Supports both Qdrant Cloud for managed hosting and local Qdrant instances for self-hosted deployments. The minimalist design philosophy keeps the interface clean and predictable, making it easy for AI models to use effectively.
✅ Best for
AI developers who want a simple, reliable semantic memory layer with minimal configuration
⏭️ Skip if
You need full vector database management capabilities like collection creation or batch operations
💡 Use cases
- Giving AI assistants persistent semantic memory across multiple conversations
- Building knowledge bases that AI can search by meaning rather than exact keywords
- Storing and retrieving project context, notes, or decisions for development workflows
- Creating personal AI memory systems that remember user preferences and history
👍 Pros
- ✓ Official Qdrant project with active maintenance and community support
- ✓ Intentionally simple two-tool interface — easy for any LLM to use correctly
- ✓ Works with both Qdrant Cloud (managed) and local instances (self-hosted)
- ✓ Docker support for quick local Qdrant setup
👎 Cons
- ✗ Only two tools — no advanced collection management or batch operations
- ✗ Requires a running Qdrant instance (cloud or local) as a dependency
- ✗ No built-in embedding configuration — relies on server-side defaults
🔧 Exposed tools (2 tools)
| Tool | Category | Description |
|---|---|---|
| qdrant_find | search | Perform semantic search across stored information |
| qdrant_store | storage | Store information with metadata in Qdrant for semantic retrieval |
⚡ Installation
Prerequisites:
- • python v3.10+
- • Qdrant Cloud API key optional (for cloud-hosted instances)
Check Claude Code documentation to configure this MCP server.
💡 Tips & tricks
For the quickest start, use Qdrant Cloud's free tier — no local setup needed. For local development, run Qdrant via Docker: docker run -p 6333:6333 qdrant/qdrant. The qdrant_store tool accepts metadata alongside the text, so include context like timestamps or categories to improve retrieval quality.
🔄 Alternatives
Quick info
- Author
- Qdrant
- License
- Apache-2.0
- Runtime
- Python 3.10+
- Transport
- stdio
- Category
- database
- Difficulty
- Beginner
- Self-hostable
- ✅
- Auth
- —
- Docker
- 🐳 Docker available
- Version
- latest
- Updated
- Feb 28, 2026
Client compatibility
- ✅ Claude Code
- ✅ Cursor
- ✅ VS Code Copilot
- ❓ Gemini CLI
- ✅ Windsurf
- ❓ Cline
- ❓ JetBrains AI
- ❓ Warp
Platforms
🍎 macOS 🐧 Linux 🪟 Windows