About OCR2Excel
OCR2Excel is a small browser tool that turns table photos and screenshots into editable spreadsheets. We built it for people who receive grids as images in chat, email, or paper and need a real .xlsx without installing desktop OCR suites.
The product philosophy is simple: keep recognition on the device, keep the first screen light, and keep the review step honest. Yellow low-confidence cells exist because machines misread characters — we would rather show uncertainty than pretend every digit is perfect.
What you can do today
On the homepage converter you can drop PNG, JPG, or WEBP images (up to ten per run), paste a screenshot, or upload a PDF (up to 10 pages), review a rebuilt grid, and download .xlsx, .csv, .docx or .txt. A multi-image run becomes one sheet per file; a PDF becomes a single continuous sheet in document order. Guides explain step-by-step conversion, OCR behavior, receipts, statements, handwriting limits, and what to do when Excel’s Data from Picture control is unavailable.
There is no account system and no sign-up wall. Inside the free limits you open the site, convert, and leave with a file. Above those limits an optional one-time unlock raises the batch ceilings; it is described on the pricing page and never blocks the free tier.
Local OCR ethos
Many free converters upload images to a cloud queue. We self-host the OCR engine files and run them with WebAssembly inside your tab after you click Convert. That design choice is deliberate for receipts, statements, and internal price lists that should not travel to a stranger’s server just to become a spreadsheet.
The recognition engine is PaddleOCR’s PP-OCRv6 (tiny), executed through ONNX Runtime Web as a single-threaded WebAssembly build. Images and scanned PDF pages go through it; pages that already carry selectable text are read directly from the PDF instead of being re-recognized, which is both faster and exact. Tesseract stays available as a fallback for browsers without WebAssembly SIMD.
Local processing is not a magic shield. You still choose what to photograph and where to store the downloaded workbook. It does remove a large class of “mystery upload” risk that worries finance and operations teams.
Open-source components
Nothing here is a black box, and nothing is fetched from a third-party CDN at runtime — every file is served from this domain. The engine files total roughly 17 MB the first time you convert; your browser caches them afterwards. The converter is built on other people’s work, and we would rather name it:
- PaddleOCR / PP-OCRv6 detection and recognition models — Apache License 2.0 (Baidu).
- paddleocr.js (ONNX Web wrapper) — MIT.
- ONNX Runtime Web — MIT (Microsoft).
- pdf.js — Apache License 2.0 (Mozilla).
- Tesseract.js and the Tesseract LSTM cores — Apache License 2.0.
- SheetJS Community Edition — Apache License 2.0.
Trademarks and project names belong to their respective owners. Listing them describes the technical stack; it does not imply endorsement, sponsorship, or affiliation by those projects or companies.
Contact
Questions, corrections, or a file format you want supported? Write to support@ocr2excel.com. The inbox is read by the person who maintains the converter, usually within a few business days.
For how-to questions the on-site guides usually answer faster, and Privacy and Terms explain how data and disclaimers are framed. Reporting a conversion that came out wrong is welcome — a short description of the file type is more useful than the file itself, and please do not send documents with personal data.
If you want to try the product, start on the homepage converter. For a structured walkthrough, read the guide on how to convert image to excel and return here when you want the company context.
OCR2Excel is built and operated by Xi’an Hongren Huizhi Software Co., Ltd.(西安鸿仁汇智软件有限公司), as a focused tool site with one job: image tables to spreadsheets in the browser. Registered address: Room 12626, Building 1, Laosanjie Shouzuo Building, No. 2 Weiyang Road, Weiyang District, Xi’an, Shaanxi, China. We will expand formats and languages carefully rather than bolting on unrelated features that bloat the first paint.
Thank you for trying the tool and for reading the limits as well as the promises. Honest docs keep expectations aligned — especially around handwriting and faded thermal receipts.
Independence
OCR2Excel is not affiliated with, endorsed by, or sponsored by Microsoft or Google. Excel, Google Sheets and the OCR engines named above are the property of their respective owners and are referenced here only to describe compatibility and the technical stack.