AI and the SDK

The Maps SDK for JavaScript has AI capabilities at two levels: one for the developers building with it, and one for the end-users of the applications they build.

For developers — AI coding assistance

Building with any SDK is faster when your coding agent understands the library. The SDK ships a /tomtom-maps-sdk-js skill that loads SDK-specific knowledge directly into AI coding agents like Claude Code, Cursor, GitHub Copilot, and Windsurf.

With the skill installed, your agent knows the right module patterns, coordinate conventions ([longitude, latitude] — always), service APIs, and common pitfalls — so it writes idiomatic SDK code on the first try rather than inventing its own.

Install it once:

1
npx skills add tomtom-international/maps-sdk-js --skill tomtom-maps-sdk-js

Then invoke it with a topic:

1
/tomtom-maps-sdk-js map-setup
2
/tomtom-maps-sdk-js routing
3
/tomtom-maps-sdk-js traffic

Learn more about AI coding with the SDK →

For end-users — conversational map control

The @tomtom-org/maps-sdk-plugin-ai-agent plugin turns a TomTom map into a conversational interface. Users can type or speak natural language — “Route from Berlin to Munich avoiding tolls”, “Show coffee shops near me”, “What’s the traffic like on this route?” — and the agent takes care of the geocoding, routing, and map display.

The plugin is headless: you bring your own LLM (any Vercel AI SDK provider) and chat UI. It ships 30+ built-in tools covering location, routing, places, traffic, and map control, with an intent classifier that keeps token usage lean.

1
import { createMapAgent } from '@tomtom-org/maps-sdk-plugin-ai-agent';
2
import { openai } from '@ai-sdk/openai';
3

4
const agent = createMapAgent(map, { model: openai('gpt-4o') });

Learn more about the AI Map Agent plugin →

Summary