GUIDE · OCR UPGRADE NOT LIVE
How do I extract a table from an image — and what works before OCR ships?
Extracting a table from an image (screenshot, phone photo, or scan) requires OCR and careful grid repair — and on PdfTableKit that image OCR path is not live yet. This page is a labeled guide. For text-layer PDFs, use the current extractor today.
Status callout
Image / scan table OCR is not available in this version. Do not treat PdfTableKit as an image-table tool until OCR ships and this page is updated with a dated capability note.
Why images need OCR
Screenshots and phone photos are pixels, not a text layer. Without OCR, there are no characters to place into cells. Deskew and contrast help later — they do not invent missing text today.
Prep checklist for future OCR
Shoot square-on, fill the frame, avoid glare, and keep columns aligned. Clean captures will matter when OCR goes live.
Why proofreading money columns matters
OCR and heuristics misread digits. Always check totals and decimal separators against the original capture.
Use the text PDF path now
If you have a text-layer PDF of the same table, extract it with the live tool on the home page. Photographed PDF pages usually behave like images, not text PDFs.
Questions
Can I upload a PNG table today?
Not as a supported extraction tool. The product shows a clear unsupported notice for image uploads on the text-PDF tool.
Why publish this page then?
Demand exists for image table extraction. We answer honestly and route text-PDF jobs to the real tool instead of faking OCR.
Will OCR be free?
When OCR ships, OCR Pack is listed at $19 / 40 and must stay labeled until live. Text packs ($9 / $19) are for text PDFs. Checkout coming soon.
Is a photographed PDF page a “text PDF”?
Usually no — photos of pages behave like images/scans.
Handwritten tables in images?
Out of scope even for the planned OCR upgrade framing in MVP non-goals.
When will this page become a tool page?
After OCR ships and the capability note is approved — until then it stays a guide.