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

Q: How do I hook up Langfuse (LLM Tracing & Evals) to Vercel AI SDK?

Install the @ai-sdk/mcp package and initialize the client using createMCPClient. You pass the tools returned by the client directly into generateText or streamText to let your model query telemetry data.

Q: Do I need to manually close the connection in my serverless routes?

You must call mcpClient.close() once your stream finishes. Leaving connections open will hang your serverless execution. The MCP protocol handles the rest.

Stream Langfuse telemetry and prompt data directly into your Vercel AI SDK frontend in real-time using our MCP Server.

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 Vercel AI SDK

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

Ask AI about this MCP

ChatGPT

Claude

Perplexity

Stream production telemetry to Vercel AI SDK UIs

When your production UI streams responses, you need to know what's happening under the hood. Connecting this MCP Server to your Vercel AI SDK setup lets your frontend trigger telemetry checks on the fly. You don't have to wait for a background job to finish before seeing how your model behaved. Use `get_trace` to pull the complete nested graph of any execution step and show it to your users or developers instantly. If a generation goes sideways, `get_observation` pulls the exact span or generation context so you can debug the UI state without digging through raw logs.

Run dynamic prompt evaluations in Next.js

Hardcoding prompts in your Next.js API routes is a bad idea. This server integration lets your application pull active prompt versions directly from your registry using `list_prompts` during the generation lifecycle. Once the user interacts with the UI, you can write human feedback directly back to the database. Call `create_score` to link star ratings or custom evaluation metrics to the active trace without leaving your edge function context.

Track daily costs and session health

Keeping an eye on API spend is a full-time job when you're streaming millions of tokens. The `get_daily_metrics` tool pulls aggregated latency and rolled-up USD costs directly into your admin dashboard. You can also group user journeys by calling `list_sessions` to trace how a single user interacted with your app across multiple API calls. It gives you a clear picture of user engagement without complex database joins.

Setup guide

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

Prerequisites

Node.js 18+ and a TypeScript project
ai + @modelcontextprotocol/sdk packages
Active Vinkius subscription with a valid endpoint token

1

Install dependencies

Run npm install ai @modelcontextprotocol/sdk plus your preferred model provider (e.g. @ai-sdk/openai).
2

Create the Streamable HTTP transport

Use StreamableHTTPClientTransport with your Vinkius endpoint URL. Replace [YOUR_TOKEN_HERE] with your token from cloud.vinkius.com.
3

Discover and use tools

Call mcpClient.tools() to auto-discover all Langfuse (LLM Tracing & Evals) tools. Pass them directly to generateText() or streamText() — no manual schema definitions needed.
4

Works with any model provider

Swap openai("gpt-4o") for any AI SDK provider — Anthropic, Google, Mistral. The MCP tools work identically across all supported models.

index.ts

import { experimental_createMCPClient as createMCPClient } from "ai";
import { StreamableHTTPClientTransport } from "@modelcontextprotocol/sdk/client/streamableHttp";
import { generateText } from "ai";
import { openai } from "@ai-sdk/openai";

const transport = new StreamableHTTPClientTransport(
  new URL("https://edge.vinkius.com/[YOUR_TOKEN_HERE]/mcp")
);

const mcpClient = await createMCPClient({ transport });
const tools = await mcpClient.tools();

const { text } = await generateText({
  model: openai("gpt-4o"),
  tools,
  prompt: "List recent Langfuse (LLM Tracing & Evals) transactions",
});

console.log(text);
await mcpClient.close();

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 Vercel AI SDK

Install the `@ai-sdk/mcp` package and initialize the client using `createMCPClient`. You pass the tools returned by the client directly into `generateText` or `streamText` to let your model query telemetry data.

Yes. The server runs in a secure V8 isolate hosted by Vinkius, which connects over standard HTTP. Your edge functions make quick, lightweight calls without carrying heavy SDK dependencies.

When a stream cuts off or errors out, your application can call `list_observations` to inspect the exact generation parameters. This lets you pinpoint if the issue was a model timeout or a parser error in your frontend.

You must call `mcpClient.close()` once your stream finishes. Leaving connections open will hang your serverless execution. The MCP protocol handles the rest.

Your raw LLM traces and prompt templates are transmitted securely through an encrypted V8 sandbox on Vinkius. No telemetry data is stored on our proxy servers; everything goes straight to your self-hosted or cloud database instance.

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