Location AI

Mapbox Location AI products support location-centered agentic workflows. It includes the Mapbox MCP Server, the Mapbox DevKit MCP Server, MapGPT, the Mapbox Feedback Agent, and the Mapbox Location Agent.

The Mapbox MCP Server provides a unified endpoint for AI systems to access the Mapbox platform of geospatial data and location technology services. The Server equips AI agents to perform tasks such as geocoding, route planning, and point-of-interest (POI) lookup through natural-language interfaces, simplifying complex geospatial workflows into a single, standardized protocol.

The hosted version of the Mapbox MCP Server is best for developers who want instant access to Mapbox services without managing infrastructure.

The local version of the Mapbox MCP Server is best for teams that require full control over data, security, and infrastructure. The MCP Server can run within your own environment for custom configuration and compliance needs.

The Mapbox DevKit MCP Server equips AI coding tools with direct Mapbox access, so they can create styles, format data, and generate tokens on demand – speeding up development and boosting creativity. Developers can install the DevKit locally or connect to the hosted version to integrate with their preferred AI coding assistants. Claude Code integration is available out of the box today, with support for additional tools like Copilot and Cursor in development.

The Mapbox MCP Server is designed to equip AI agents that interact with end users, giving them access to geospatial reasoning and geographic data. The DevKit, meanwhile, supports developers who are building with Mapbox services by enabling them to connect AI coding assistants to the Mapbox platform to improve productivity and automate repetitive project tasks. 

In short: the Mapbox MCP Server helps AIs like agents or LLMs understand location context while the DevKit helps developers build AI experiences faster with Mapbox.

MapGPT is a conversational copilot AI assistant designed for use specifically in automotive in-vehicle systems. Learn more at https://www.mapbox.com/mapgpt.

The Mapbox Feedback Agent enables voice-based feedback from end users so they can easily report updates or errors in map data or navigation instructions. AI analysis enriches feedback with context and categorization to speed review and response by service teams. Learn more at https://www.mapbox.com/feedback-agent.

The Mapbox Location Agent is a demonstration environment of how the Mapbox MCP Server makes maps conversational – turning user questions into intelligent, location-based insights. Request access to the private preview >>

The Mapbox Location Agent is a demonstration AI-powered map assistant that showcases how the Mapbox Model Context Protocol (MCP) Server connects large language models (LLMs) with geospatial data.

Yes, both MCP Servers require a valid Mapbox access token, available within your Mapbox account. It is free to sign up for a Mapbox account.

No, you can try the Mapbox Location Agent private preview demo without a Mapbox account. A free Mapbox account is only required if you want to build your own implementation or use the Mapbox MCP Server directly.

Use of the Mapbox MCP Server and DevKit themselves do not incur costs, but the API calls generated while using them may incur costs if the volume exceeds the monthly free thresholds (see Mapbox Pricing for details).

The Mapbox MCP Server and DevKit are both available to any Mapbox account holder. Get started by reviewing the documentation for the Mapbox MCP Server . To join the Location Agent private preview, submit an access request.