Langfuse MCP Server
Access and manage your Langfuse prompt library directly from your AI agent interface.
Data last scanned today · Reviewed today
Overview
The Langfuse MCP Server integrates Langfuse Prompt Management into Model Context Protocol-compatible clients. It enables developers to discover and retrieve prompts stored within their Langfuse account, facilitating consistent prompt usage across various AI assistant platforms. The server supports two primary methods for access: the native MCP Prompts specification for prompt discovery and retrieval, and a set of tools that replicate this functionality for clients lacking native prompt support. It retrieves prompts labeled as 'production' and compiles them with specified variables to ensure accurate usage.
Our verdict
This server carries a high-risk rating of 26/100, primarily because it is a community-maintained project that has not been editorially reviewed. Given that the last commit was 568 days ago and critical security aspects like authentication and read-only mode have not been reviewed, users should exercise caution. This tool is best suited for experimental environments rather than sensitive production workflows where maintainer activity and security audits are required.
- Uses only three direct dependencies to minimize supply chain footprint
- Licensed under the permissive MIT license
- Provides documented support for MCP prompt discovery and retrieval
- Project has not received a commit in over 560 days
- Maintained by the community rather than the official vendor
- Security architecture for authentication remains unreviewed
Setup
{
"mcpServers": {
"mcp-server-langfuse": {
"args": [
"-y",
"mcp-server-langfuse"
],
"command": "npx"
}
}
}Paste into ~/Library/Application Support/Claude/claude_desktop_config.json (macOS) or %APPDATA%\Claude\claude_desktop_config.json (Windows), then restart Claude Desktop. Some servers require API keys or tokens as environment variables — check the project's README and add an "env" object if needed.
- No read-only mode — running this grants full read/write access.
- No built-in authentication — anyone with access to this config can invoke every tool it exposes.
Tools
| Tool | Description | Risk |
|---|---|---|
| get-prompts | List available prompts | low |
| get-prompt | Retrieve and compile a specific prompt | low |
Compatibility
| Client | Local | Docker | Remote | Read-only |
|---|---|---|---|---|
| Claude Desktop | ||||
| Cursor | ||||
| VS Code | ||||
| Windsurf | ||||
| ChatGPT |
Frequently asked questions
›Which Langfuse prompts are accessible via this server?
Only prompts that have a production label assigned to them in Langfuse are returned by the server.
›What authentication information is required to run the server?
You must provide your LANGFUSE_PUBLIC_KEY and LANGFUSE_SECRET_KEY as environment variables when configuring the server.
›Are prompt arguments fully documented when retrieved?
No, all arguments are assumed to be optional and do not include descriptions because variables currently lack a specification within Langfuse.
›Why are there both MCP prompts and tools in this server?
The server exports tools as an alternative to the standard MCP prompts capability to ensure compatibility with MCP clients that do not support the prompt specification.
Changelog
- Improvedv0.0.2-rc.0
2025-02-16
- Newv0.0.1
2025-02-15
Badge
Maintain this server? Add the live badge to your README.