qdrant-mcp-local

Name: qdrant-mcp-local
Author: hirokita117

@hirokita117

代码仓库

关于 qdrant-mcp-local

ローカル環境でQdrantとMCP-server-qdrantを簡単に立ち上げるためのDocker Compose設定

#qdrant #docker #mcp-server

基本信息

分类

数据库

运行时

shell

传输方式

stdio

发布者

hirokita117

配置

使用下面的配置,将此服务器添加到你的 MCP 客户端。

{
  "mcpServers": {
    "qdrant-mcp-local": {
      "command": "docker",
      "args": [
        "compose",
        "up",
        "-d"
      ]
    }
  }
}

工具

未检测到工具

工具是从 README 中自动提取的。维护者可以在 ## Tools 标题下列出工具,即可填充这部分内容。

概览

What is qdrant-mcp-local?

A Docker Compose configuration that launches Qdrant vector search engine and MCP-server-qdrant (a Model Context Protocol server for Qdrant) in a local environment.

How to use qdrant-mcp-local?

Clone the repository, create a data/ directory, then run docker compose up -d. Qdrant is available at http://localhost:6333 and the MCP server at http://localhost:8000. Stop with docker compose down. Configuration can be customized via a .env file.

Key features of qdrant-mcp-local

Single-command local setup of Qdrant and MCP-server-qdrant
Persistent vector data stored in ./data/ directory
SSE transport endpoint at http://localhost:8000/sse
Configurable environment variables via .env file
Includes debug script for troubleshooting

Use cases of qdrant-mcp-local

Local development of AI applications requiring vector search
Testing and prototyping MCP integrations with Qdrant
Running a self-contained Qdrant + MCP stack for experimentation
Evaluating Qdrant features without cloud dependencies

FAQ from qdrant-mcp-local

What are the prerequisites?

Docker and Docker Compose V2 are required.

How do I customize the configuration?

Edit the .env file to change environment variables (e.g., port numbers).

Where is vector data stored?

Data is persisted in the ./data/ directory, so it survives container restarts.

What endpoints are exposed?

Qdrant at http://localhost:6333 and the MCP SSE endpoint at http://localhost:8000/sse.

How do I troubleshoot if containers don't start?

Run the debug.sh script, check logs with docker compose logs mcp-server, or rebuild with docker compose build --no-cache. Common issues include port conflicts, insufficient disk space, or a stopped Docker daemon.