Image to Excel Converter
Convert image tables to Excel data online free — extract tabular data from photos using OCR. Nothing leaves your device — fast and private.
This tool turns a picture of a table into a spreadsheet. Drop in a screenshot or photo, and OCR reads the text while a clustering pass rebuilds the rows and columns. You get an editable grid to fix any misread cells, then download a real .xlsx workbook or a .csv. Nothing is uploaded.
Why retyping tables is the wrong answer
Data arrives in stubbornly uneditable forms: a paper bank statement, a printed price list, a supplier catalogue, a screenshot from a web portal that offers no export button. Retyping it is slow, and worse, it is where errors enter. A single transposed digit in a column of figures produces a total that is wrong in a way nobody notices. Extraction plus a quick visual check is both faster and more accurate than manual entry, particularly for the long numeric columns that make up statements and invoices.
- Bank statements and receipts into a spreadsheet for bookkeeping or a tax return.
- Price lists and catalogues from a supplier PDF or screenshot into something sortable.
- Printed reports and survey results into rows you can pivot and chart.
- Marksheets and result tables into a sheet for calculating averages.
How the rows and columns are reconstructed
OCR alone returns a flat run of words, which is useless for a table. The structure comes from the coordinates. Every recognised word carries a bounding box, and the tool uses those in two passes. Rows come from vertical position: words whose boxes sit within a threshold derived from the average glyph height on the page are grouped onto the same line. Columns come from horizontal gaps: the tool measures average character width, then treats any gap wider than roughly three characters as a column boundary and records where each column starts. Words are then dropped into the matching cell, and empty rows and columns are pruned. This is why gaps matter more than gridlines — a table with generous spacing between columns extracts well even without borders, while a tightly packed one can merge two columns into one.
What extracts well, and what does not
| Input | Expected result |
|---|---|
| Screenshot of a spreadsheet or web table | Very good — clean text, even spacing |
| Sharp scan of a printed, ruled table | Very good, with occasional character fixes |
| Phone photo of a printed statement | Workable if shot square-on in even light |
| Tightly packed columns, no gaps | Columns may merge — widen or crop and retry |
| Merged cells or nested headers | Flattened; fix in the grid or in Excel |
| Cells wrapping onto two lines | Split across rows; needs manual repair |
| Handwritten tables | Poor — typing is usually faster |
Getting a cleaner extraction
- Prefer a screenshot to a photo. No perspective, no shadow, perfectly sharp text.
- Crop to the table itself, removing page furniture, logos, and surrounding paragraphs.
- Shoot square-on if you must photograph paper, and keep shadows off the page.
- Feed it resolution. Small text is the main cause of character errors.
- Check the numeric columns first in the grid, since a wrong digit matters far more than a wrong letter.
Privacy for the documents this tool is actually used on
The tables people convert are rarely trivial: bank and credit card statements, payroll lists, invoices, supplier contracts, medical billing. Uploading any of those to a cloud OCR service hands a third party both the image and every figure extracted from it, and leaves you inheriting their retention policy. Because recognition and spreadsheet generation both run locally here, there is no upload, no server-side copy, and nothing to chase for deletion afterwards.
Related tools
If you want plain text rather than a grid, or you need a language other than English,Image to Text handles eight. For a table inside a text-based PDF,PDF to Excel reads the embedded text positions and avoids OCR error altogether. For a scanned PDF, make it searchable with OCR PDF or export the pages usingPDF to JPG and bring one here.
Frequently Asked Questions
How do I convert a photo or screenshot of a table into Excel?
Drop a JPG, PNG, WebP, or BMP image of the table into the tool. OCR reads the words and their positions, a clustering pass groups them into rows and columns, and you get an editable grid to correct before downloading as .xlsx or .csv.Is my bank statement or invoice uploaded to a server?
No. Recognition runs through Tesseract compiled to WebAssembly and the spreadsheet is built by SheetJS, both inside your browser. The image and the extracted numbers never leave your device, which is the entire reason to use this rather than a cloud converter.Which image formats are supported?
JPG, PNG, WebP, and BMP, one image at a time. A screenshot is usually better input than a photograph because it has no perspective distortion, even lighting, and perfectly sharp text.How accurate is table extraction?
Clean screenshots of spreadsheets and high-contrast printed tables come back close to right, with occasional character errors. Ruled tables with clear gaps between columns work best. Merged cells, nested headers, and wrapped multi-line cells are where the structure breaks down.Which language does the OCR read?
English only in this tool, because the column-detection pass is tuned around English word spacing. For other languages use our Image to Text tool, which offers eight, then paste the result into a spreadsheet and split the columns manually.What is the difference between the .xlsx and .csv downloads?
The .xlsx file is a real Excel workbook with the data on a sheet named Extracted. The .csv is plain text that opens anywhere, including Google Sheets and Apple Numbers, and is written with a byte-order mark so Excel reads accented characters correctly.Can it read handwritten tables?
Not reliably. The engine is trained on printed and digital type, so neat block capitals sometimes work while cursive generally does not. Handwritten ledgers are usually faster to type manually than to correct afterwards.Can I extract a table from a PDF?
Convert the page to an image with our PDF to JPG tool first, then bring it here. For a text-based PDF, our PDF to Excel tool is the better route because it reads the embedded text positions directly and skips OCR errors entirely.