Visual Report Demo

This self-authored page is a public demo input for doc7. It deliberately combines a chart, table, formula, workflow, and screenshot-like UI state in one document.

Start with the before/after showcase, then compare the raw PDF with the Markdown output.

Create the raster input:

rsvg-convert -o input.png source.svg

Process it with any configured OpenAI-compatible vision model:

doc7 read input.png -o output

See output.md for one real model result. It preserves the visible text and table, writes the formula as LaTeX, and describes the chart, workflow, and UI state with [Visual] blocks. Results vary with the configured model.

input.pdf contains the same page exported as PDF. output-pdf.md shows the result after the PDF renderer path; its key values, table rows, formula, and visual relationships remain consistent with the PNG result.