LIVE
Mistral Makes Large 4 Available as Open Weights29/10/26 · Mistral AI|Gemini 4 Carbon tops the Artificial Analysis Intelligence Index v529/10/26 · Google|Mistral Large 4 now available on Hugging Face29/10/26 · Mistral AI|Google Introduces "Gemini 4 Carbon" as Next Step in Its Model Lineup28/10/26 · Google|DeepMind unveils Genie 4, a new generation of world model27/10/26 · Google DeepMind|Anthropic files IPO prospectus, with risk disclosures taking center stage26/10/26 · Anthropic|Robot policy trained only in simulation matches top real-world systems26/10/26|OpenAI unveils GPT-6.1 Terra, a model built for computer use26/10/26 · OpenAI|New benchmark checks whether LLMs actually test their temporal medical claims29/10/26|An ODE-based sampling method speeds up discrete diffusion models29/10/26|OpenAI revises its Model Spec with five guiding values for AI behavior29/10/26 · OpenAI|Inside Proton's internal clash over its AI pivot29/10/26 · Proton|Mistral Makes Large 4 Available as Open Weights29/10/26 · Mistral AI|Gemini 4 Carbon tops the Artificial Analysis Intelligence Index v529/10/26 · Google|Mistral Large 4 now available on Hugging Face29/10/26 · Mistral AI|Google Introduces "Gemini 4 Carbon" as Next Step in Its Model Lineup28/10/26 · Google|DeepMind unveils Genie 4, a new generation of world model27/10/26 · Google DeepMind|Anthropic files IPO prospectus, with risk disclosures taking center stage26/10/26 · Anthropic|Robot policy trained only in simulation matches top real-world systems26/10/26|OpenAI unveils GPT-6.1 Terra, a model built for computer use26/10/26 · OpenAI|New benchmark checks whether LLMs actually test their temporal medical claims29/10/26|An ODE-based sampling method speeds up discrete diffusion models29/10/26|OpenAI revises its Model Spec with five guiding values for AI behavior29/10/26 · OpenAI|Inside Proton's internal clash over its AI pivot29/10/26 · Proton|
MCPNewofficial publisher

Qdrant MCP Server

MCP : Standardises the connection between an AI application and systems that expose tools or data · requires any-agent

Official Qdrant MCP server: gives Claude, Cursor, or any compatible agent semantic memory, with two actions, storing information and retrieving it by meaning.

The simplest way to connect a Qdrant database to an assistant so it can remember notes, decisions, or code snippets from one session to the next. It requires a Qdrant instance already in place, either local or in the cloud.

Official site GitHub Docs
Overall score
5.9 /10
Security: one caveat, see the card
Utility
6.0
Reliability
6.0
Security
One caveat
Ease
Technical · Technical

Security reflects where data goes, requested permissions and incidents on record at the verification date. It is not a guarantee of zero risk: check the deployment mode and the connections you enable.

What it does, concretely

An AI assistant has no memory between sessions: each conversation starts from scratch. Qdrant MCP Server solves this by providing a persistent semantic memory layer accessible from any MCP-compatible agent.

In practice, the server exposes two actions to the agent: storing information in a Qdrant collection, and retrieving information by the meaning of a natural language query. The default embedding model is sentence-transformers/all-MiniLM-L6-v2, provided via FastEmbed. The collection is created automatically if it does not yet exist.

This tool is aimed at developers using Claude Desktop, Cursor, Windsurf, VS Code, or Claude Code who want their assistant to remember notes, architecture decisions, or code snippets from one session to the next. It can also serve as a semantic search engine over a code snippet database.

It is not suited for someone looking for a turnkey solution without command-line work: configuration is done through environment variables and requires a Qdrant instance already available, locally or in the cloud.

What it offers

  • ·Store information in a Qdrant collection via the qdrant-store action, with optional JSON metadata.
  • ·Semantic search by natural language via the qdrant-find action, returning up to 10 results by default (configurable).
  • ·Automatic collection creation if it does not exist at the time of the first store operation.
  • ·Read-only mode activatable via the QDRANT_READ_ONLY variable, which disables the write action.
  • ·Customizable tool descriptions via environment variables, to adapt agent behavior to specific contexts (e.g. code search).
  • ·Support for three transport protocols: stdio for local use, SSE and streamable-HTTP for remote or shared connections.
  • ·Compatible with Claude Desktop, Cursor, Windsurf, VS Code, and Claude Code, as well as any standard MCP client.
  • ·Can be deployed via uvx, Docker, or Smithery, with no prior installation needed when using uvx.

In practice

Price
Free
Commercial use
Commercial use: yes
Available on
cli
Host
any-agent
Licence
Apache 2.0
FR interface
No

Security and data

Rating
Security: one caveat, see the card
Your data
Local or cloud depending on use
Maintenance
Active

Links and repository checked on 29 October 2026 · added on 29 October 2026

How to install Qdrant MCP Server

1. Make sure uvx is installed (via the uv package manager). 2. Have a Qdrant instance available: locally on port 6333 or via Qdrant Cloud. 3. Set the environment variables QDRANT_URL (or QDRANT_LOCAL_PATH), COLLECTION_NAME, and if needed, QDRANT_API_KEY. 4. Start the server with: uvx mcp-server-qdrant (add --transport sse for remote access). 5. Add the corresponding configuration to claude_desktop_config.json, Cursor/VS Code MCP settings, or via the claude mcp add command for Claude Code. 6. Use a dedicated collection for the agent to avoid exposing your production data.
uvx mcp-server-qdrant

Official documentation ↗

In the same category

Whole category Feed (memory, context) →