A screenshot or photo of a table is just pixels: there is no text to copy, let alone columns. This tool reads it with OCR (the Tesseract engine) and then rebuilds the table. It looks for the ruled lines between cells, and where a table has none it finds the columns from the gaps that run straight down it, the same way our PDF to Excel converter does for PDFs. The result appears as an editable grid beside the picture, so you can fix any misread cell before saving it as an Excel file or CSV, or copying it straight into a spreadsheet.
Before reading, the picture is cleaned up for OCR: small screenshot text is enlarged, a tilted photo is straightened, uneven light and dark-mode colours are evened out, and the table's lines are removed so they cannot be mistaken for letters. Recognition runs in your browser, so the picture is never uploaded. It is free to use, with no account and no upload limit, and it runs entirely in your browser: nothing is uploaded, and while the page stays open, a feature you have already used keeps working offline.
- How accurate is converting a table image to Excel?
- On our 13 test pictures of printed tables (screenshots, scans and phone photos), from about three cells in four to every cell came out exactly right before any correction, depending on the picture, and every table had its rows and columns in the right place. Clear, large text does best; very small screenshot text, blur, glare and unusual fonts bring more mistakes, mostly a digit misread or a decimal point or comma lost. The tool puts back many lost points and commas by looking at the rest of the column, and highlights the cells it is unsure of or has corrected, which in our tests included about three in five of the cells it got wrong. So the highlighted cells are the place to start, but check every number that matters against the picture before you rely on it.
- Can it read a handwritten table?
- No. The recognition engine is trained on printed and on-screen text, and handwriting comes out as nonsense. A table typed into a form and then printed or photographed is fine.
- Why are some cells highlighted?
- A highlighted cell was either read with low confidence or changed by the tool: for example 1299 in a column of prices like 12.50 becomes 12.99, and an O in a column of numbers becomes a 0. These are usually right, but they are the cells worth checking first. Hover over one to see why it is marked and, for a corrected cell, what was read; select it and press Put back what was read to undo the correction.
- Will the numbers from my picture work in Excel formulas?
- Yes. Amounts with thousands separators, either decimal separator, currency signs, percentages and negatives in brackets are saved as numbers, formatted the way they were printed. Dates written so they cannot be misread, such as 2026-03-31 or 12 Jan 2026, are saved as dates; 03/04/2026-style dates are saved as dates only when the column shows which number is the day, and are otherwise kept as text. A cell starting with = is always saved as text, never as a formula.
- Can I convert a scanned PDF with this?
- Yes: drop the PDF in and each page is read as a picture. If the PDF has real text in it, such as a statement downloaded from a bank, PDF to Excel is the better choice, as it reads the text exactly instead of recognising it.
- Is my picture uploaded to a server?
- No. The picture is read in your browser on your own computer. The only download is the recognition engine and the language data, the first time you use each language; the language data comes from tessdata.projectnaptha.com, which learns which language you chose but never sees your picture.