How to Use the Langfuse (LLM Tracing & Evals) MCP in Claude Code

Q: Can I pull raw trace graphs using Claude Code and Langfuse (LLM Tracing & Evals)?

Yes, Claude Code calls gettrace to retrieve the complete nested graph for any trace ID you specify. This allows you to inspect deep execution trees directly within your terminal window.

Q: What is the setup command to add this MCP Server to Claude Code?

Run claude mcp add --transport http langfuse-llm-tracing-evals-mcp -- in your terminal. Note that all transport flags must be placed before the server name.

Q: How does Claude Code handle prompt management with Langfuse (LLM Tracing & Evals)?

The terminal agent runs listprompts to fetch all actively managed prompt templates and their versions. This lets you inspect system instructions directly from your command-line interface.

Run terminal commands to audit LLM costs, fetch production traces, and manage prompt templates directly from your shell.

See Vinkius in Action

Works with every AI agent you already use

…and any MCP-compatible client

MCP Servers - Free for Subscribers

Connect Langfuse (LLM Tracing & Evals) MCP to Claude Code

Create your Vinkius account to connect Langfuse (LLM Tracing & Evals) to Claude Code and route execution through our secure gateway. The platform manages server hosting, runtime updates, and security layers. Configuration requires no manual server provisioning.

GDPR Free for Subscribers

Setup Langfuse (LLM Tracing & Evals) with Claude Code

Ask AI about this MCP

ChatGPT

Claude

Perplexity

Audit production LLM metrics from your terminal

To monitor system health, Claude Code executes `get_daily_metrics` to retrieve aggregated USD costs and latency statistics directly in your shell. This allows SREs and backend engineers to check operational performance without opening a browser. If the metrics reveal a sudden spike in latency, you can pipe the output of this MCP call to other CLI utilities or have the agent run `list_traces` to isolate the slow requests. It keeps your monitoring workflow fast and terminal-centric.

Pipe trace data and observations into shell scripts

Inspecting raw telemetry is simple because the CLI uses `get_observation` to output explicit span and generation details as clean JSON. You can easily feed this data into grep, jq, or custom bash scripts to automate post-mortem reports. When deeper debugging is required, running the MCP tool `list_observations` lets you scan across thousands of traces to find pattern failures. The terminal client parses these payloads in milliseconds to find bad LLM outputs.

Manage prompts via this terminal-based MCP Server

Keeping track of prompt versions in production is straightforward since the agent uses `list_prompts` to extract deployed templates. This command lets you quickly verify which version of a prompt is currently active in your environment. You can also use `create_observation` to log new events or spans directly from automated CLI test runs. It makes integration testing of prompt chains easy to manage from any CI/CD pipeline.

Setup guide

Set up Langfuse (LLM Tracing & Evals) MCP in Claude Code

Prerequisites

Claude Code CLI installed (npm install -g @anthropic-ai/claude-code)
Active Vinkius subscription with a valid endpoint token

1

Run the add command

Open your terminal and run the command shown on the right. Replace [YOUR_TOKEN_HERE] with your endpoint token from cloud.vinkius.com. Use --scope user to make it available across all projects.
2

Verify the connection

Start a Claude Code session and type /mcp to list connected servers. You should see langfuse-llm-tracing-evals-mcp with a green status indicator.
3

Start using tools

Ask Claude Code something like "Check my latest Langfuse (LLM Tracing & Evals) transactions." It will automatically discover and invoke the available Langfuse (LLM Tracing & Evals) tools.

Terminal

claude mcp add --transport http langfuse-llm-tracing-evals-mcp https://edge.vinkius.com/[YOUR_TOKEN_HERE]/mcp

Get your connection token →

Prerequisites

Claude Code CLI installed
Active Vinkius subscription with a valid endpoint token

1

Open the config file

Create or edit .mcp.json in your project root for project-level scope, or ~/.claude.json for user-level scope.
2

Add the Langfuse (LLM Tracing & Evals) MCP

Paste the JSON snippet shown on the right into the mcpServers object. Replace [YOUR_TOKEN_HERE] with your endpoint token from cloud.vinkius.com.
3

Restart Claude Code

Start a new Claude Code session. Type /mcp to confirm the server is connected. The tools will be automatically available in your conversation.

.mcp.json

{
  "mcpServers": {
    "langfuse-llm-tracing-evals-mcp": {
      "type": "url",
      "url": "https://edge.vinkius.com/[YOUR_TOKEN_HERE]/mcp"
    }
  }
}

Get your connection token →

Independent Platform Disclaimer: Vinkius is an independent platform and is not affiliated with, endorsed by, sponsored by, verified by, or otherwise authorized by Langfuse. All third-party trademarks, logos, and brand names are the property of their respective owners. Their use on this website is strictly for informational purposes to identify service compatibility and interoperability.

Why Choose Vinkius

Vinkius connects your tools to AI with real-time monitoring and automatic cost savings — all from one dashboard.

Connect Langfuse (LLM Tracing & Evals) now

Real-time monitoring

Live

visibility into every interaction

Connect your favorite tools to your AI and see exactly what's happening — every request, every response, in real time.

Built-in savings

60%

lower AI costs

Vinkius compresses data between your apps and your AI automatically. Lower bills every month — no configuration required.

Single dashboard

One

place for every integration

Every tool your AI connects to, managed from a single screen. One account, complete control.

Common questions about Langfuse (LLM Tracing & Evals) MCP in Claude Code

You can ask the agent to check your daily spend, and it will execute `get_daily_metrics` to pull the rolled-up USD costs. The terminal output can then be piped directly to other CLI tools or saved to a file.

Yes, Claude Code calls `get_trace` to retrieve the complete nested graph for any trace ID you specify. This allows you to inspect deep execution trees directly within your terminal window.

Run `claude mcp add --transport http langfuse-llm-tracing-evals-mcp -- ` in your terminal. Note that all transport flags must be placed before the server name.

The terminal agent runs `list_prompts` to fetch all actively managed prompt templates and their versions. This lets you inspect system instructions directly from your command-line interface.

Absolutely. Your LLM traces, observations, scores, and prompt templates are processed through an ephemeral, zero-trust MCP sandbox. The server does not persist your data, keeping your production logs completely private.

Use it with your favorite AI tools

Connect this server to Cursor, Claude, VS Code, and more.

OpenAI Agents SDK sdk-python

Google ADK sdk-python

Pydantic AI sdk-python

Vercel AI SDK sdk-typescript