How to Use the Mapbox MCP in Pydantic AI

Q: How does Pydantic AI ensure coordinates from the Mapbox geocode tool are valid?

When your agent calls geocode, Pydantic AI receives the JSON response and validates it against a model. It checks that the coordinate fields exist and are the correct type, like floats. If validation fails, it stops execution with an error, preventing your Pydantic AI agent from using bad data.

Q: Can I use Pydantic AI with Mapbox to create a route and get a typed Python object back?

Yes, that's exactly what it's for. When your agent calls getdirections, Pydantic AI parses the response into a structured object. You work with an instance of a class with attributes like .duration and .distance, not a raw dictionary you have to parse yourself.

Q: What's the difference between MCPToolset and MCPServerHTTP for Pydantic AI?

MCPToolset is the modern, correct way to connect to this MCP Server. It's a unified interface that handles different transports. MCPServerHTTP is deprecated, so you should use MCPToolset("http://...") for any new projects.

Build correct, type-safe location agents with Pydantic AI and Mapbox. Get validated routes, geocoded addresses, and reliable map data.

See Vinkius in Action

Works with every AI agent you already use

…and any MCP-compatible client

MCP Servers - Free for Subscribers

Connect Mapbox MCP to Pydantic AI

Create your Vinkius account to connect Mapbox 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 Mapbox with Pydantic AI

Ask AI about this MCP

ChatGPT

Claude

Perplexity

Prevent Bad Location Data

Pydantic AI validates every response from the Mapbox server against a Pydantic model. If the `geocode` tool returns a malformed address or `get_elevation` returns a non-numeric value, your agent raises a `ValidationError` on the spot. This means no silent data corruption. Your application won't get a `None` where it expects coordinates or a string where it expects a float. You build with confidence, knowing the location data your agent uses is always correct and properly structured.

Work with Any LLM, Get Consistent Results

Pydantic AI is model-agnostic, so you can use OpenAI, Anthropic, Gemini, or a local model to run your agent. The code for calling the Mapbox tools doesn't change. Your agent's capabilities are decoupled from the brain powering it. Using the `MCPToolset` standardizes how your agent calls the `get_directions` or `search_nearby` tools. This lets you swap the underlying LLM to find the best balance of cost and performance, all without rewriting your tool-handling logic.

Define Your Own Data Structures

While the MCP Server provides a default output structure, Pydantic AI lets you define your own models for the results. You can create a custom `Route` model that only includes the `distance` and `duration` fields from the `get_directions` tool's output. The framework automatically parses the full response and populates your specific object. This keeps your business logic clean and separated from the raw API response of tools like `get_distance_matrix` or `get_isochrone`.

Setup guide

Set up Mapbox 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": {
        "mapbox-mcp": {
            "url": "https://edge.vinkius.com/[YOUR_TOKEN_HERE]/mcp"
        }
    }
})

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

result = await agent.run("List recent Mapbox 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 Mapbox. 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 Mapbox 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 Mapbox MCP in Pydantic AI

When your agent calls `geocode`, Pydantic AI receives the JSON response and validates it against a model. It checks that the coordinate fields exist and are the correct type, like floats. If validation fails, it stops execution with an error, preventing your Pydantic AI agent from using bad data.

Yes, that's exactly what it's for. When your agent calls `get_directions`, Pydantic AI parses the response into a structured object. You work with an instance of a class with attributes like `.duration` and `.distance`, not a raw dictionary you have to parse yourself.

It will fail loudly and safely. If the Mapbox API returns a new structure that doesn't match your Pydantic model, you'll get a `ValidationError`. This is a feature, not a bug—it tells you immediately that the data contract has changed so you can update your models.

`MCPToolset` is the modern, correct way to connect to this MCP Server. It's a unified interface that handles different transports. `MCPServerHTTP` is deprecated, so you should use `MCPToolset("http://...")` for any new projects.

Only the data you explicitly save yourself. Your requests, containing things like coordinate pairs for `get_elevation` or search queries for `search_nearby`, pass through Vinkius's secure, ephemeral environment to Mapbox. Vinkius does not persist any of your request or response data.

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