AI-Generated Maps: From a Topic to a Full Animated Timeline

The Next Step: Letting AI Build the Map

Creating an animated historical map on Moving Polygons involves real research: identifying the right time periods, tracing borders at each stage, and assembling everything into a coherent timeline. It's rewarding work, but it's also time-consuming. That naturally led us to a question: what if an AI could handle the heavy lifting?

We've been experimenting with using large language models to generate complete maps from nothing more than a topic. Give it a subject — an empire, an era, a conflict — and the model conducts the historical research, determines the key territorial changes, and produces a fully structured map with polygons, timestamps, and labels. Everything that previously required hours of manual work, automated in minutes.

Fast First Drafts, Not Perfect History

To be clear: AI-generated historical research is not inherently more accurate than doing it by hand. Language models can hallucinate details, merge time periods, or simplify complex territorial boundaries. What they excel at is speed. An LLM can produce a strong first version of a map in a fraction of the time it would take to build one from scratch.

That's where Moving Polygons' editor comes in. Once the AI has generated the initial map, you can open it in the editor and refine every detail: adjust polygon shapes for geographic accuracy, correct dates, add or remove timeline entries, and fix any errors the model introduced. The combination of AI-generated drafts and human refinement is where the real power lies — a complete animated historical map that didn't exist before, created from nothing more than a topic.

How It Works Today

We've been building a fine-tuned AI agent that can take a topic, ask clarifying questions about scope and time period, and then generate the full map data. The vision is to have this integrated directly into Moving Polygons — type a topic, answer a few questions, and get a complete animated map.

In practice, the direct API integration isn't quite there yet. The research phase requires the model to process and synthesize substantial amounts of historical information, and the resulting output — detailed polygon coordinates for every territorial change — can be too large for a single API response to handle reliably. It works, but not consistently enough for a smooth user experience.

The JSON Import Workaround

In the meantime, there's a practical workaround that works surprisingly well. Moving Polygons supports importing and exporting maps as JSON files. This means you can use any LLM — ChatGPT, Claude, or any other model — to generate map data, and then import it directly into the platform.

Here's the workflow: first, export an existing map as JSON from your dashboard. This gives you the exact format that Moving Polygons expects. Then, provide that JSON structure to your LLM of choice along with the historical topic you want to map. The model will use the format as a template and fill it with researched territorial data. Finally, save the output as a .json file and import it into Moving Polygons using the GeoJSON import feature.

• Export any of your existing maps as JSON from the dashboard to get the correct format
• Provide the JSON structure and your chosen topic to an LLM (e.g., ChatGPT or Claude)
• Ask the model to research the territorial changes and generate polygon data in the same format
• Save the LLM's output as a .json file
• Import the file into Moving Polygons via the GeoJSON import on your dashboard or in the editor
• Refine the result in the editor for historical accuracy

It requires a bit of back-and-forth with the model to get clean, valid output — but once you have a working JSON file, the import is instant. And because the editor gives you full control over every polygon and timestamp, you can correct any inaccuracies the model introduced.

What's Coming

We're actively working on making the AI generation more reliable and integrating it directly into the platform. The goal is a seamless experience: describe what you want to visualize, and Moving Polygons builds it for you. The underlying technology works — it's the reliability and output size constraints that need solving.

In the meantime, the JSON import workflow is a genuinely useful path to creating maps that would otherwise take hours of manual research and polygon drawing. If you try it out, we'd love to hear about your experience.