# langfuse/mcp-server-langfuse [Health: Active]

**Category:** 🗣️ Conversational AI  
**Repository:** https://github.com/langfuse/mcp-server-langfuse  
**GitHub Stars:** 176  
**Views:** 4  
**Installs:** 0  
**Upvotes:** 0  
**Directory Page:** https://allmcps.com/mcp/langfuse-mcp-server-langfuse

## Description
MCP server to access and manage LLM application prompts created with Langfuse) Prompt Management.

## Tools
Capabilities this server exposes over MCP:

- **submitFeedback** — Submit explicit user-approved feedback to the Langfuse team.
Before calling, ask permission and show the exact payload.
If the user wants a reply, ask them to include their email address in the feedback text; only use an address they explicitly provide and show it in the exact payload preview.
Do not include secrets, credentials, customer/project data, trace payloads, or unrelated context; the only contact detail to include is an explicitly provided reply email.
- **searchLangfuseDocs** — Semantic search (RAG) over the Langfuse documentation. Use this whenever the user asks a broader question that cannot be answered by a specific single page. Returns a concise answer synthesized from relevant docs. The raw provider response is included in _meta. Prefer this before guessing. If a specific page is needed call getLangfuseDocsPage first.
- **getLangfuseDocsPage** — Fetch the raw Markdown for a single Langfuse docs page. Accepts a docs path (e.g., /docs/observability/overview) or a full https://langfuse.com URL. Returns the exact Markdown (may include front matter). Use when you need a specific page content (Integration, Features, API, etc.) or code samples. Prefer searchLangfuseDocs for broader questions where there is not one specific page about it. Changelog pages (/changelog/...) are historical release notes: use them only to confirm that a feature exists or when it shipped, not as an implementation reference. Their code samples reflect the SDK/API at release time and may be outdated, so for implementation follow the current docs and the API/SDK reference instead.
- **getLangfuseOverview** — Get a high-level, machine-readable index by downloading https://langfuse.com/llms.txt. Use this at the start of a session when needed to discover key docs endpoints or to seed follow-up calls to searchLangfuseDocs or getLangfuseDocsPage. Returns the plain text contents of llms.txt. Avoid repeated calls within the same session.

## Claude Desktop Quick Installation
No confirmed install details yet, so no config is published. Follow the setup steps in the repository above.

## Documentation

## What langfuse/mcp-server-langfuse MCP server does

langfuse/mcp-server-langfuse MCP server connects an MCP client to prompts stored in Langfuse Prompt Management. It supports prompt discovery and retrieval for both text and chat prompts, exposing the results through the MCP Prompts specification and through tools for clients without prompt-capability support.

The server is focused on reading and compiling prompts rather than editing them. Listing returns available prompt names and their arguments. Retrieval selects a prompt by name and applies the variables supplied by the client.

## How it works

The MCP prompt interface provides `prompts/list` for discovery and `prompts/get` for retrieving a prompt. The server also exposes matching `get-prompts` and `get-prompt` tools. The tool form of retrieval requires a `name` and accepts an optional JSON `arguments` object containing prompt variables.

List operations can use a cursor for pagination. To determine the arguments associated with each prompt, the server fetches prompts individually in the background. Retrieved Langfuse text and chat prompts are transformed into MCP prompt objects, and supplied variables are used during compilation.

Only prompts carrying the `production` label are returned. The implementation treats all discovered arguments as optional, and argument descriptions are not included because Langfuse variables do not provide the specification required for them.

## Setup and configuration

The repository is built locally with npm. Run `npm install`, then `npm run build`; the resulting entry point is `build/index.js`. An MCP client should launch that file with Node.js.

The server requires three environment variables: `LANGFUSE_PUBLIC_KEY`, `LANGFUSE_SECRET_KEY`, and `LANGFUSE_BASEURL`. The first two authenticate requests to Langfuse, while the base URL identifies the Langfuse endpoint.

For Claude Desktop, add a server entry whose command is `node`, whose argument points to the built `build/index.js`, and whose environment contains the Langfuse credentials and base URL. Cursor can also run the server as a command with the same environment variables and Node.js entry point.

## Tools and capabilities

- `get-prompts` lists available production prompts and supports an optional pagination cursor.
- `get-prompt` retrieves and compiles one prompt using its required `name` and optional JSON variables.
- `prompts/list` provides MCP-native prompt discovery with pagination support.
- `prompts/get` provides MCP-native prompt retrieval for text and chat prompts.

The tool endpoints mirror the MCP prompt operations, which gives clients an alternative when they do not implement the prompt capability directly.

## Limitations and notes

langfuse/mcp-server-langfuse MCP server is described as a work in progress. Prompt listing is less efficient because the server retrieves prompts individually to extract their arguments. Returned arguments are marked optional, and no variable descriptions are supplied. Prompt visibility is also restricted to entries labeled `production`; prompts without that label will not appear in discovery or retrieval results.

_Full upstream README: https://allmcps.com/mcp/langfuse-mcp-server-langfuse/readme_