Image to Markdown Converter (OCR)
Convert an image to text and markdown with OCR in your browser. Upload a screenshot or photo and get editable markdown with headings and lists.
Drop a screenshot or photo here
or paste an image from the clipboard, or click to choose a file
PNG, JPG, WEBP, GIF. OCR runs in your browser; the first run downloads about 10 MB of language data.
What Is an Image to Text Converter?
An image to text converter reads the characters in a picture with optical character recognition (OCR) and returns them as editable text. This version goes one step further and produces markdown: the recognised lines are grouped into headings, paragraphs, bullet lists, numbered lists and simple tables, so a screenshot of documentation, a slide or a receipt becomes something you can paste into a README, a note or a ticket.
Recognition runs in your browser with tesseract.js. Nothing is uploaded, which makes it suitable for internal screenshots and documents you cannot send to a cloud OCR service.
How to Convert a Screenshot to Markdown
- Drop an image on the left pane, click to choose a file, or paste a screenshot from the clipboard.
- Select the language of the text. English is the default.
- Click Extract text. The progress bar shows the language download and the recognition steps.
- Read the markdown in the middle pane and fix any misread characters; the preview on the right updates as you type.
- Toggle Detect tables or Join wrapped lines if the layout was guessed wrong.
- Copy the markdown or download it as a .md file.
OCR to Markdown Formatting Rules
| Recognised text looks like | Markdown output |
|---|---|
| A line in ALL CAPS on its own | # Heading |
| A short line without a full stop, blank line after it | ## Heading |
| Lines starting with •, -, – or * | - bullet |
| Lines starting with 1. or 1) | 1. item |
| Consecutive lines with the same number of 2-space gaps | | table | row | |
| Wrapped lines of a paragraph, hyphen at line end | Joined into one paragraph, hyphenation removed |
Getting Better Results from Image to Markdown
OCR quality depends on the pixels it receives. Capture screenshots at 100% zoom or larger, prefer dark text on a light background, and crop away toolbars and sidebars so the columns detector only sees the content. For code, the monospace font recognises well but indentation is approximate; paste the result into the code formatter afterwards. Very large photos are slower because the whole image is scanned; a 1,500 pixel wide screenshot usually finishes in a few seconds after the first run.
When to Use an Image to Text Tool
Reach for the image to text converter when the source is only available as pixels: a slide from a recorded talk, an error dialog, a table in a scanned PDF, a whiteboard photo after a meeting, or a screenshot of a document someone sent instead of the file. The markdown output slots into the other tools on this site: run it through the table formatter, the markdown to DOCX converter or the markdown editor to finish the job.
Frequently Asked Questions
How accurate is image to text OCR in the browser?
For clean screenshots of text at normal size, Tesseract reads almost every word correctly. Accuracy drops with photos taken at an angle, low contrast, small fonts under about 20 pixels tall and handwriting. Zooming the screenshot before capturing it helps more than any setting.
Is my image uploaded to a server?
No. The image to text step runs entirely in your browser with tesseract.js compiled to WebAssembly. The only network request is the one-time download of the language data (about 10 MB for English), which your browser caches.
How does screenshot to markdown detect headings and lists?
OCR returns plain lines, so the formatter applies rules: a short line on its own without a full stop becomes a heading, ALL CAPS lines become top-level headings, lines starting with bullets or numbers become lists, and columns separated by two or more spaces become a table. You can edit the result before copying it.
Which languages are supported?
English, German, French, Spanish, Portuguese, Italian, Dutch, Japanese, Simplified Chinese, Korean and Russian. Pick the language before you click Extract text; each language downloads its own data file on first use.
Can I convert a photo of a printed page?
Yes, if the page is flat, evenly lit and fills most of the frame. Crop out margins and other objects first. For multi-page scans, use the PDF to Markdown tool on a PDF instead of photographing each page.