Server facts
better-qdrant-mcp-server
Project
Tech
- Runtime / language
- node
- Transports
- stdio
- Remote available
- No
- Authentication
- Not specified
Activity
- GitHub stars
- 4
- Last commit
- Data as of
Tags
- better qdrant
- mcp server
- vector database
Overview
Document ingestion and similarity search sit alongside collection management in better-qdrant-mcp-server. Its four tools let users inspect available collections, add documents to a chosen collection, search for related documents, and delete a collection. The document-ingestion example takes a file path and exposes settings for chunk size, chunk overlap, and embedding service. Search accepts a query, a target collection, an embedding-service choice, and a result limit. These operations serve people working with Qdrant data through an MCP client, including those using the documented Claude configuration.
Embedding options include OpenAI and OpenRouter with API keys, plus local models through Ollama and FastEmbed. Environment variables supply the Qdrant connection details and embedding-provider settings, with support for a project-root .env file. The Node.js server communicates over stdio and requires Node.js 18.0.0 or later and an active Qdrant instance, whether hosted locally or remotely. Distribution is through npm, with npx also documented as a way to launch it. The project uses the MIT license.
better-qdrant-mcp-server setup notes
- better-qdrant-mcp-server records stdio only; confirm a network transport upstream before planning remote access.