Connecting to the world of AI agents: Introducing TomTom Model Context Protocol Server

Today, TomTom releases the alpha version of TomTom Model Context Protocol (MCP) server, a solution that enables AI agents access to TomTom data and APIs, bringing more context and spatial awareness to different AI use cases. We explore the use of AI solutions at TomTom and speak to our product team about TomTom MCP Server.

Technological shifts are how businesses evolve. Smartphones transform the way we navigate the world. The Internet of Things connects everything around us. Big data enables us to see and understand patterns at scale. Through all of these innovations, TomTom has continued to adapt and often lead industry.

Today, we're witnessing another major transformation: AI. It's changing how people search, plan and build digital experiences. It’s about a lot more than just chatbots — AI is solving dynamic real-world problems, from itinerary planning to driving assistance.

This isn’t TomTom’s first encounter with intelligent tech. In fact, the company has been working with big data and machine learning for decades. To date, the mapmaker has archived 2.8 trillion kilometers of global distance data, gathered over 48 trillion GPS points and continues to gather 3.5 billion kilometers of live driving data every day, among billions of other data points — big data can’t get any bigger than global map data. Using this vast amount of data, TomTom has built sophisticated analytical models, used deep machine learning and consistently delivered useful location-based products. TomTom Co-Founder and CMO, Corinne Vigreux often mentions,

“

For three decades TomTom has consistently ridden wave after wave of technological breakthroughs. We’ve been at the forefront adapting and innovating, be it making our maps with deep machine learning or improving traffic with analytical algorithms. We were an AI-tech company long before AI became a buzzword.

However, innovation isn’t all plain sailing. When it comes to connecting AI models with data sources, we are seeing some growing pains. The way AI models like LLMs access data from various sources can define how well the AI performs. We’ve all experienced moments where AI hallucinates, produces incorrect information and fails to provide an accurate response. To solve this, AI needs to be integrated with (real-time) data sources, in essence vastly improving its corpus of knowledge. However, such integration can create fragmentation and inefficiencies between the AI models and the sources, as each platform’s integration can be unique. A standardized, secure, and developer-friendly access to real-world data is needed.

Which brings us to today’s announcement. We spoke with Senior Product Manager, Ruben Woelders to explain what is happening.

Ruben Woelders: TomTom is launching the alpha version of its Model Context Protocol (MCP) Server — a powerful new way to enable AI agents to connect to TomTom APIs and data. Our mapping, geolocation and traffic data is known to be some of the best out there and with the MCP server, AI models such as Claude, ChatGPT, Llama, Mistral and many others can tap into real-time location data, making them significantly more context-aware and capable in travel, logistics, retail and other location-driven use cases. The product is in alpha at the moment, which means that it’s at an initial development stage where we invite developers to explore, share insights, and help shape the future of the product. The accuracy of the output will improve as we build future iterations.

Ruben Woelders: Model Context Protocol (MCP) links AI models to data sources like APIs, files and databases. The protocol has seen increasing adoption since it was launched by Anthropic in November 2024. TomTom MCP Server is the way MCP-enabled AI applications get access to TomTom’s high-quality maps, location or traffic data. Think of it as the translator that allows an AI agent to ask meaningful questions about the real world and get accurate, up-to-date answers from TomTom. For example, think about a retail analyst who might use TomTom data accessible through an LLM such as Claude to make decisions that support their business plans. Here the LLM could effectively access not only TomTom data but also other sources to offer a contextual solution to the retail analyst.

See below for examples of how Claude gets real-time context from TomTom’s MCP server:

With our MCP server, developers can use TomTom APIs simply by asking for what they need in natural language, without integrating APIs one-by-one. And the power is not just limited to chat interfaces like Claude, but any application could be powered with our MCP server.

Ruben Woelders: A context-aware AI agent understands not just the question you're asking, but also the surrounding situation: your location, past interactions, preferences and current conditions. It’s like talking to someone who knows you and your environment well — so instead of vague responses, you get accurate, personalized and relevant suggestions. For example, “Find me a route home that avoids traffic and passes a pharmacy” is only possible if the AI understands both your intent and your real-world location context.

The real-world location context comes from the data created by organizations such as TomTom.

Ruben Woelders: AI without context can’t deliver accuracy. LLMs are powerful, but they’re not inherently connected to real-time data sources. Ask a typical LLM, “What’s the fastest route to work right now?” and you’ll likely get a generic or outdated answer. With the MCP Server, AI agents can access live data from TomTom, including maps, traffic, routing and points of interest — leading to more intelligent, useful, and reliable outputs.

MCP is a common protocol to streamline and standardize access. Developers no longer need to build custom connectors for each application, which could affect the accuracy of responses. Instead, they can simply use the MCP server and immediately enable AI to interface with TomTom’s world-class data.

Ruben Woelders: We foresee a wide range of users and invite developers to try the alpha MCP server for the following use cases:

• Developers building AI agents or applications that need geographic, routing, traffic or POI data

• Location analytics businesses or teams that need accurate spatial data to make informed decisions, for example; driving times, identifying nearby competitors and assessing parking availability.

• Public sector organizations: urban planners using AI for city services can now incorporate real-time mobility data.

• Fleet & logistics companies that want to leverage real-world traffic data and commercial vehicle routing This is just the beginning. We expect to see MCP used across industries, from transportation and retail to smart cities and autonomous vehicles.

Ruben Woelders: The alpha TomTom MCP Server is now available on GitHub, complete with open-source code, documentation and setup guides. Developers can start building immediately and connect their LLMs or AI agents to the real-world using TomTom’s data streams.

You can explore it now at:

👉 developer.tomtom.com/tomtom-mcp

👉 github.com/tomtom-international/tomtom-mcp

We will continue to work on the product and provide documentation and tutorials as we go further. For more technical information read our developer blog.

As AI becomes the new interface for how we interact with the digital world, location context will be essential. With MCP, TomTom is making it easier for AI agents to see, understand and respond to the world as it really is.