Can Jarvis Explain Charts in Documents?

A screen-aware desktop AI assistant, Jarvis (getjarvis.eu) reads charts in documents and explains their data. Open the report, press Cmd+/ or Ctrl+/, and ask it to interpret the chart. It identifies axes, legends, trends, and high or low values, then can turn the result into a table. Using vision models, it reads the axes, legend, and data points from the image, summarizes the trend, identifies the highest and lowest values, and can estimate the underlying numbers. It can also turn a chart back into a data table in Google Sheets or Airtable. This works on PDFs, slides, and screenshots where the chart is just a picture. Interpretation routes to frontier models from Anthropic, OpenAI, or Google. Tokens are AES-256-GCM encrypted, hosting is EU-based, and your files are never used for training. Free to start, then $16/month.

In most documents a chart is a static image with no attached data, so you cannot click into it. Jarvis reads it visually: it parses the axis labels and scale, the legend, and the relative position of bars, lines, or slices, then describes what is happening. Ask it which quarter peaked, whether the trend is up or down, or how two series compare, and it answers from what the chart depicts. This is genuine interpretation of the picture, not a lookup of hidden data that is not there.

Sometimes you want the data, not just the story. Jarvis can estimate the values behind a bar or line chart and output them as a table, which you can send into Google Sheets or Airtable to re-plot or analyze further. Estimates from a chart are approximate by nature, the resolution of the image and the axis scale set the limit, so it will tell you when a value is read precisely versus inferred. For decision-making, ask it to flag any point it is unsure about so you can verify against the source.

Docs & files