How to Use the Weights & Biases MCP in Pydantic AI

Guarantee correctness with Weights & Biases using Pydantic AI.

See Vinkius in Action

Works with every AI agent you already use

…and any MCP-compatible client

MCP Servers - Free for Subscribers

Connect Weights & Biases MCP to Pydantic AI

Create your Vinkius account to connect Weights & Biases 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 Weights & Biases with Pydantic AI

Ask AI about this MCP

ChatGPT

Claude

Perplexity

Get Guaranteed Run Details

When you call `get_run_details`, the MCP Server validates every single piece of data—metrics, configs, summaries. Your agent gets a guaranteed, type-safe object back; no messy dictionaries. This is huge for correctness. If W&B sends unexpected fields or types, your Pydantic AI client fails loudly with an error instead of letting the system process corrupted data.

List Artifacts Safely

The `list_project_artifacts` tool provides a list whose structure is enforced by Pydantic models. You always know exactly what fields (like name and type) you're working with. This eliminates the risk of silent data corruption that plagues agents relying on generic JSON responses. Correctness first, every time.

Check Sweep Parameters

Using `list_project_sweeps`, your agent pulls hyperparameter boundaries into a clean, typed structure. You don't have to write custom parsing logic for W&B data. Because the output is validated against Pydantic models, you can trust that the list of sweeps and their parameters are structured exactly as expected by your production code.

Setup guide

Set up Weights & Biases 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": {
        "weights-biases-mcp": {
            "url": "https://edge.vinkius.com/[YOUR_TOKEN_HERE]/mcp"
        }
    }
})

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

result = await agent.run("List recent Weights & Biases 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 Weights & Biases. 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 Weights & Biases 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 Weights & Biases MCP in Pydantic AI

The MCP Server wraps W&B functionality, guaranteeing that every response—whether a run detail or an artifact list—conforms to your defined Python types. The agent never has to guess the data shape.

Yes, while this specific client focuses on structured reads, the underlying MCP Server supports listing projects via `list_wandb_projects`. The key benefit is that even these lists are validated upon return.

You can access run details, artifact metadata, and sweep configurations. Every field you read—like a metric value or a project name—is guaranteed to be the correct type (string, float, integer).

Yes, that's its whole point. It validates the API response against a defined schema at runtime. If W&B changes its API format unexpectedly, your agent fails with an explicit validation error.

The server touches structured run metrics and configs. Since the output is strictly typed, it makes handling sensitive model performance data much more predictable and safer in your production pipeline.

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