Headings From Type Size
Text noticeably larger than the body of the page becomes a # heading at the matching level. Size ratio drives the decision, so the same rules work on a phone photo and a 600 DPI scan.
Convert Image to Markdown with private in-browser OCR. Turn screenshots, scans, PNG, JPG, WebP, and GIF files into editable headings, lists, and tables without an upload.
The Image to Markdown converter reads the text in a screenshot, scan, or photo and writes it back out as Markdown — headings, lists, and tables included — using OCR that runs inside your own browser rather than on a server.

Image to Markdown recognises words and positions, rebuilds document structure, and keeps the result adjustable.
"Follow-up actions" 18pt bold y=557
"Item" "Owner" "Due" 11pt y=599
"Clear drainage outlet" … x=372, 723, 938## Follow-up actions
| Item | Owner | Due |
| --- | --- | --- |
| Clear drainage outlet | Maintenance | 14 days |Image to Markdown uses OCR geometry to preserve headings, lists, tables, languages, and readable paragraphs.
Text noticeably larger than the body of the page becomes a # heading at the matching level. Size ratio drives the decision, so the same rules work on a phone photo and a 600 DPI scan.
Rows whose words land in the same vertical bands are rebuilt as a GFM pipe table — the hardest part of Image to Markdown conversion, and the one worth checking.
Bullet glyphs and 1. / a) markers become nested list items. Lines broken by the page margin are joined back into paragraphs, words hyphenated across a break are rejoined, and a wide vertical gap starts a new block.
English plus Chinese, Japanese, Korean, Russian, Arabic and the major European languages, in mixed-script pairs so a bilingual document is read in one pass. Switching the language reads every open image again in it.
The converter reports its own average confidence and how many uncertain words it discarded, so you know whether to trust the result or rescan at a higher resolution.
Recognition runs locally on your own machine. Unlike hosted OCR services, your screenshot, contract scan, or whiteboard photo is never transmitted to or stored on a server.
Convert Image to Markdown from a screenshot, scan, or photo without an account or upload.
Paste a screenshot or add PNG, JPG, WebP, and GIF files up to 15 MB each; every image joins the conversion queue.
Image to Markdown starts automatically and reads the queue with one shared OCR engine. Choose the language before or after adding images.
Compare the preview with the source, adjust heading, list, and table detection, then copy Markdown, download .md, or export the batch as a ZIP.
Answers about OCR accuracy, privacy, languages, tables, code, handwriting, and caching.
Paste a screenshot or add an image, then turn unselectable text into editable Markdown.