Vizzy

Vizzy is a browser-based data visualization tool built around a simple idea: large language models are good at reasoning about data but poor at drawing it. Vizzy splits the job. You describe your data to an LLM, or hand it a raw export, and ask for a structured chart or map definition. Then you paste that JSON into Vizzy, which takes care of rendering it as a real, interactive visualization.

Because the output is declarative JSON rather than a specific chart type, the same approach covers bar charts, line charts, scatter plots, tables and geographic maps over a base layer. Vizzy can also ingest the underlying data directly in JSON, CSV or XML, so you are not limited to hand-crafted specs; uploading a dataset and a matching spec is often the fastest path from raw numbers to something you can look at.

The project markets itself as open source and rapid, targeting people who want a quick visual answer rather than a full dashboard build. Its homepage ships with a set of ready-made demo datasets — scripts from the TV show The Office, changes in same-sex marriage legislation across U.S. states over time, U.S. state obesity rates with shape geometry, CDC cause-of-death coverage versus news mentions, and a U.S. national parks dataset with geographic and activity detail. Those examples double as templates showing what kinds of data the tool handles well.