Table Screenshot to Markdown
Paste or drop a screenshot of any table and get clean Markdown table syntax. OCR runs 100% locally in your browser — your image is never uploaded. Works with Chinese and English tables.
1. Screenshot
2. Markdown result (editable)
Tip: for mostly-Chinese tables keep "Chinese" selected — mixed-language OCR can misread CJK near symbols like "->". Merged cells and diagonal headers are not supported; please review the result.
FAQ
Is my screenshot uploaded to a server?
No. The OCR engine (Tesseract.js) runs entirely in your browser via WebAssembly. Your image never leaves your device.
Why does the first run take longer?
The OCR engine and language data (a few MB) are downloaded from a CDN on first use and cached by your browser afterwards.
What kind of tables work best?
Clear, high-resolution screenshots with left-aligned columns and visible spacing between columns. Merged cells and diagonal headers cannot be detected — fix those cells manually after recognition.