How to Use the Langfuse (LLM Tracing & Evals) MCP in Pydantic AI

Q: How do I install Langfuse (LLM Tracing & Evals) in Pydantic AI?

Install the slim package with pip install "pydantic-ai-slimmcp". Initialize MCPToolset with your Vinkius HTTP URL and pass it to the toolsets array in your Agent. Avoid the deprecated MCPServerHTTP class.

Q: Does Pydantic AI validate Langfuse (LLM Tracing & Evals) responses?

Yes. Every payload from tools like getobservation or listprompts runs through strict Pydantic models. Any structural mismatch triggers an immediate runtime exception.

Q: How do I retrieve raw observation objects?

The agent calls listobservations to pull events spanning across multiple traces. You get exactly the span context you request.

Bring the Langfuse (LLM Tracing & Evals) MCP Server to Pydantic AI for type-safe telemetry extraction and strict evaluation auditing.

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 Pydantic AI

Create your Vinkius account to connect Langfuse (LLM Tracing & Evals) to Pydantic AI 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 Pydantic AI

Ask AI about this MCP

ChatGPT

Claude

Perplexity

Extract Type-Safe Traces

`get_trace` and `list_traces` pull complete telemetry data that Pydantic AI validates instantly. If the API returns a missing field or malformed nested graph, the framework throws a loud validation error. You never deal with silent data corruption in your logs. Correctness drives this integration. Your agent relies on predictable data structures to analyze past behavior. Forcing strict schemas on the returned JSON ensures your downstream evaluation logic never breaks unexpectedly.

Enforce Quality with Langfuse MCP Server

`create_score` attaches explicit quality or cost algorithms directly to an observation. Your agent can list these metrics via `list_scores` to audit its own performance over time. Every score format is validated before the agent processes it. Model-agnostic systems require standardized evaluation. Whether you route requests to commercial APIs or a local model, the scoring mechanism remains identical. You maintain a single source of truth for output quality.

Audit Daily Metrics and Sessions

`get_daily_metrics` generates aggregated USD cost and latency statistics for your agent to review. `list_sessions` provides the high-level user entities that encapsulate those traces. The framework guarantees the types of these financial and timing metrics. Building reliable agents means tracking their operational footprint. You can set up automated jobs via this MCP connection that fetch these stats and fail loudly if costs exceed predefined models. Budget enforcement becomes a strict code requirement.

Setup guide

Set up Langfuse (LLM Tracing & Evals) MCP in Pydantic AI

Prerequisites

Python 3.10+ installed
pydantic-ai-slim[fastmcp] package
Active Vinkius subscription with a valid endpoint token

1

Install Pydantic AI with FastMCP
Run pip install "pydantic-ai-slim[fastmcp]". The FastMCP toolset replaces the deprecated MCPServerHTTP class with full protocol support.
2

Configure the FastMCPToolset
Pass a JSON-style config dict to FastMCPToolset with your Vinkius URL. Replace [YOUR_TOKEN_HERE] with your token from cloud.vinkius.com. Supports Streamable HTTP, SSE, and Stdio transports.
3

Create and run your agent
Pass the toolset to Agent(toolsets=[toolset]) and call agent.run(). Swap openai:gpt-4o for any supported model — Anthropic, Google, Mistral, or Groq.

agent.py

from pydantic_ai import Agent
from pydantic_ai.toolsets.fastmcp import FastMCPToolset

toolset = FastMCPToolset({
    "mcpServers": {
        "langfuse-llm-tracing-evals-mcp": {
            "url": "https://edge.vinkius.com/[YOUR_TOKEN_HERE]/mcp"
        }
    }
})

agent = Agent(
    "openai:gpt-4o",
    toolsets=[toolset],
    system_prompt="You have access to Langfuse (LLM Tracing & Evals) tools.",
)

result = await agent.run("List recent Langfuse (LLM Tracing & Evals) transactions")
print(result.output)

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 Pydantic AI

Install the slim package with `pip install "pydantic-ai-slim[mcp]"`. Initialize `MCPToolset` with your Vinkius HTTP URL and pass it to the `toolsets` array in your Agent. Avoid the deprecated `MCPServerHTTP` class.

Yes. Every payload from tools like `get_observation` or `list_prompts` runs through strict Pydantic models. Any structural mismatch triggers an immediate runtime exception.

The integration is entirely model-agnostic. Your Pydantic AI agent can track traces and log scores regardless of which underlying model generates the text.

The agent calls `list_observations` to pull events spanning across multiple traces. You get exactly the span context you request.

The Vinkius V8 Isolate Sandbox processes your requests in memory. Human feedback scores, aggregated USD costs, and raw generation texts are discarded immediately after execution. Your endpoint token is the only requirement for access.

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