Table OCR.
Run table OCR on images, scans, screenshots, or PDFs and turn visible table rows into Excel-ready CSV.
- CSV rows
- 10
- CSV columns
- 8
- Sections
- 2
Start with the file you already have.
Upload or paste the source, then check the workbook before export.
Upload or paste the source. The cleanup rule is loaded.
Sample table OCR to menu CSV
A public Wikimedia price menu photo becomes a CSV with section labels, item rows, visible prices, blank currency fields, row order, and review notes.
A public table photo becomes 10 clean CSV rows.
This sample uses a public Wikimedia Commons price menu photo for Eaton Park. The generated CSV keeps the FRUIT and CRÈME sections, item names, visible prices, blank currency cells, row order, and review-ready fields separate.
CSV preview from the public price menu photo
These rows come from the public menu photo. The CSV keeps section labels and row order, while currency stays blank because the image does not show a visible currency symbol or code.
| section | item_name | size_or_option | price | currency | row_order | review_note |
|---|---|---|---|---|---|---|
| FRUIT | Pumpkin | 13.99 | 1 | |||
| FRUIT | Apple | 14.99 | 2 | |||
| FRUIT | Blackberry | 16.99 | 3 | |||
| FRUIT | Dutch Apple | 15.99 | 4 | |||
| CRÈME | Banana Crème | 13.99 | 6 |
Source trace
The table source stays attached to every row so the CSV can be checked against the original image.
Section labels
Section labels are preserved as fields instead of being copied into item names.
Item rows
Each visible menu item becomes one row with the price copied exactly as visible.
No invented values
Fields not supported by the photo stay blank, so the CSV does not invent currency, sizes, taxes, or fees.
Plain image OCR
Plain OCR often gives text in reading order, but it does not decide which values are rows, sections, notes, or unsupported fields.
- Section headers can merge with item names
- Prices can lose row order
- Missing currency can be guessed incorrectly
Reviewable table CSV
Messy2Sheet uses the target CSV fields to preserve item rows, section labels, row order, and blank unsupported fields.
- One CSV row per visible menu item
- FRUIT and CRÈME remain section values
- Currency stays blank because the photo does not show it
What table OCR should preserve.
Table OCR should preserve visible rows, section labels, values, row order, blank unsupported fields, and uncertain values so the output can be checked.
Set the CSV fields before extraction
Set fields such as source file, section, item name, size or option, price, currency, row order, review note. Messy2Sheet uses those fields to shape the CSV output before extraction starts.
Read section labels, item rows, and visible prices together
Table OCR needs section labels, item rows, visible values, row order, and source notes, not only recognized characters. Keep blank or unsupported values visible instead of guessing them.
Check the result before downloading CSV
Compare the generated rows with the source, fix unclear values, then export CSV when the table is ready.
Use it when table image, PDF, scan, screenshot, or pasted OCR text needs structured CSV rows.
Table OCR works best when the source is readable and the target fields are clear. Keep review notes when the source layout or values need human judgment.
Best Table OCR inputs
- Readable table image, PDF, scan, screenshot, or pasted OCR text where a person can identify the important values.
- Cases where you already know the output fields: source file, section, item name, size or option, price, currency, row order, review note.
- Sources with repeated business layouts that can become a saved workflow.
- Files that need Excel or CSV output you can check before using.
Check before exporting CSV
- The source is blurry, cropped, skewed, or low contrast.
- Important values are handwritten, hidden, or missing.
- Several unrelated layouts are mixed in one upload.
- The result requires business decisions that are not visible in the source.
Start with a table ocr layout you can reuse.
For repeated files, set the output fields once, check one result, then reuse the saved layout for the next similar source.
CSV output layout
Use a target sheet when the final CSV file needs stable fields every time.
Saved Table OCR workflow
Save the instructions that handle repeated headings, wrapped text, missing fields, and source-specific cleanup rules.
- 01Keep rows in a checkable order.
- 02Split source text into the requested columns.
- 03Flag unclear cells for manual checking.
Save this table OCR layout for repeat table scans.
Start from the real file
Upload a table image, PDF, scan, screenshot, or pasted OCR text or paste source text if you already have it.
Set the fields
Use fields like source file, section, item name, size or option, price, currency, row order, review note so the generated CSV file matches the job.
Save the layout
Save the setup when the result looks right, then reuse it for similar files.
More ways to turn documents into Excel.
If this is not the file on your desk today, start with a nearby workflow and adjust the columns before running it.
Bank statement converter
For bank pages, statements, deposits, withdrawals, balances, and fee sections.
ConvertBank Statement Excel Template
Use a bank statement Excel template when the final columns need to stay consistent every month. Fill transaction rows, summary rows, balances, source pages, and review notes from a readable statement.
StartPDF Bank Statement to Excel
Convert PDF bank statements to Excel-ready CSV rows with dates, descriptions, money in, money out, balances, statement context, and review notes. Keep summary rows and daily balances separate before download.
StartPDF to CSV converter
For PDF tables, scans, screenshots, price lists, reports, and target CSV fields.
ConvertExtract Tables from PDF
Extract table sections from readable PDFs into CSV or Excel rows. Preserve table labels, source pages, NA rows, repeated headers, multiple values, row order, and review notes.
StartQuestions before you upload?
What does Table OCR do?+
Table OCR reads table images, PDFs, scans, or screenshots and exports headers, section labels, row labels, visible values, row order, source file, and review notes to Excel or CSV.
Can it preserve row order?+
Yes. Add row_order, source_file, section, item_name, price, and review_note fields when the order of rows matters.
What image quality works best?+
Readable table images with visible headers and row labels work best. Blur, skew, and glare need manual checking.
Can it leave unsupported values blank?+
Yes. If the image shows a price but no visible currency symbol or size, keep currency or size fields blank instead of inventing them.
Can it skip non-row notes?+
Yes. Ask for one row per visible item or table row, and keep availability notes, disclaimers, or captions out of the item table unless you need them as review notes.
Can I reuse table OCR fields?+
Yes. Save the layout for repeated table scans or similar reports.
Try it on one real file.
Upload the document, keep or edit the template instructions, and check the workbook before you download it.