Qdrant

Store notes and find them by meaning in a Qdrant vector collection.

Needs your setup first
★ 1.6K GitHub stars
11.4K downloads a week
Last commit 4 Sept 2026

At a glance

Two tools: store and find. Text goes in with optional metadata, gets embedded locally, and comes back when a later question is close in meaning. It works as long-term memory for an assistant or as a searchable snippet library.

It runs against a local Qdrant, Qdrant Cloud, or an on-disk path with no server at all.

Made by: Qdrant
Licence: Apache-2.0
Latest version: 0.8.1 (10 Dec 2025)
PyPI package: mcp-server-qdrant
GitHub stars: 1.6K
Downloads: 11.4K a week
Last commit: 4 Sept 2026
Tools: 2
Runs on: Your computer

Our check: It needs something our Linux test machine does not have (an account, a desktop app, a Mac or a running service), so on 26 Oct 2026 we read its tools from the source code instead of running this config. The npm package called "mcp-server-qdrant" comes from a different repository (hlos-ai/mcp-servers). Qdrant publishes this server to PyPI under that name, so it runs with uvx.

What's in the pod: its tools

We found 2 tools in its source code on 26 Oct 2026. 1 only read. The rest can change things, so keep your app's "ask before each tool call" setting on for them.

How we tested it: Read from the source of mcp-server-qdrant 0.8.1. Our sandbox could not download the embedding model it fetches on first start. qdrant-store is left out when QDRANT_READ_ONLY is true.

Changes things1 tool · Can create, edit, send or delete.

Tool	What it is for
`qdrant-store`	Saves a piece of text, with optional metadata, as a searchable memory in a Qdrant collection.

Only reads1 tool · Looks things up and changes nothing.

Tool	What it is for
`qdrant-find`	Searches the stored notes by meaning and returns the closest matches.

Grouped and described by us. The README on GitHub lists the inputs each tool takes.

Set it up

Pick your app for the exact snippet and where it goes.

Where it goes: claude_desktop_config.json (Settings → Developer → Edit Config)

Claude Desktop

{
  "mcpServers": {
    "qdrant": {
      "command": "uvx",
      "args": [
        "mcp-server-qdrant"
      ],
      "env": {
        "QDRANT_URL": "http://localhost:6333",
        "COLLECTION_NAME": "assistant-memory"
      }
    }
  }
}

Where it goes: Your terminal, in the project folder

Claude Code

claude mcp add-json qdrant '{"command":"uvx","args":["mcp-server-qdrant"],"env":{"QDRANT_URL":"http://localhost:6333","COLLECTION_NAME":"assistant-memory"}}'

Where it goes: ~/.cursor/mcp.json, or .cursor/mcp.json in a project

Cursor

{
  "mcpServers": {
    "qdrant": {
      "command": "uvx",
      "args": [
        "mcp-server-qdrant"
      ],
      "env": {
        "QDRANT_URL": "http://localhost:6333",
        "COLLECTION_NAME": "assistant-memory"
      }
    }
  }
}

Where it goes: .vscode/mcp.json in your project

VS Code

{
  "servers": {
    "qdrant": {
      "type": "stdio",
      "command": "uvx",
      "args": [
        "mcp-server-qdrant"
      ],
      "env": {
        "QDRANT_URL": "http://localhost:6333",
        "COLLECTION_NAME": "assistant-memory"
      }
    }
  }
}

Where it goes: ~/.codeium/windsurf/mcp_config.json

Windsurf

{
  "mcpServers": {
    "qdrant": {
      "command": "uvx",
      "args": [
        "mcp-server-qdrant"
      ],
      "env": {
        "QDRANT_URL": "http://localhost:6333",
        "COLLECTION_NAME": "assistant-memory"
      }
    }
  }
}

New to this? See the setup guides for Claude Code, Cursor, Claude Desktop and ChatGPT.

Settings

Names are exactly as the project spells them. Required ones are marked.

Setting	What it does
`QDRANT_URL`Environment variable	Address of a Qdrant server or Qdrant Cloud cluster. Use this or QDRANT_LOCAL_PATH, not both.`http://localhost:6333`
`QDRANT_API_KEY`Environment variable	API key for a secured or cloud Qdrant server.
`QDRANT_LOCAL_PATH`Environment variable	A folder to keep an embedded Qdrant database in, so no server is needed.`/path/to/qdrant-data`
`COLLECTION_NAME`Environment variable	The default collection. When set, the agent no longer chooses one.`my-collection`
`EMBEDDING_MODEL`Environment variable	Which FastEmbed model turns text into vectors. Defaults to sentence-transformers/all-MiniLM-L6-v2.
`QDRANT_READ_ONLY`Environment variable	Set to true to hide qdrant-store so the agent can only search.`true`
`QDRANT_SEARCH_LIMIT`Environment variable	Most results one search returns. Defaults to 10.
`TOOL_STORE_DESCRIPTION`Environment variable	Rewrites the store tool's description, for example to make it a code snippet store. TOOL_FIND_DESCRIPTION does the same for search.
`--transport`Command line flag	stdio by default, or sse or streamable-http to serve remote clients.`--transport streamable-http`

What it can touch

It can read

The Qdrant collection you point it at

It can change

Adds new points to that collection when it stores a memory
Downloads the embedding model the first time it runs

Safer setup: Set QDRANT_READ_ONLY=true when the agent only needs to look things up.

Before you start

uv (for uvx)
A Qdrant instance, or a local path for embedded storage

If it does not work

It refuses to start with both a URL and a path: Pick one: QDRANT_URL for a server, or QDRANT_LOCAL_PATH for an embedded database.
Search finds nothing in an existing collection: The collection must have been filled with the same embedding model. Vectors from another model will not match.
First call is slow: The embedding model downloads on first use. Later calls are quick.

Peas in a pod: similar servers

How Qdrant lines up against other databases and data servers on the shelf.

Server	Tools	Runs on	Licence	Stars
Qdrant	2	Your computer	Apache-2.0	1.6K
Excel MCP Server	46	Your computer	MIT	4.3K
Postgres MCP Pro	9	Your computer	MIT	3.4K
Supabase	29	Hosted	–	3K
MCP Toolbox for Databases	29	Your computer	Apache-2.0	16.7K