Extract data from images

Photos, screenshots, and scans — turn any image into structured data. ExtractFox reads photos the way a person does, so receipts, IDs, sign boards, charts, and screenshot tables all extract cleanly.

Drop a PDF or image here, or click to browse

Max 20 MB per file · PDF, PNG, JPG, WEBP, HEIC

Pro: drop up to 25 files at once for bulk extraction

What to extract from this image?

or describe it yourself

Extracting tables

Why this matters

Image-based extraction has historically meant OCR plus brittle post-processing. With a multimodal model, the layout, table boundaries, and context come for free — even on phone photos with glare or angle.

How it works

Step 1
Upload an image
PNG, JPG, WEBP, or HEIC. Phone photos and screenshots both fine.
Step 2
Describe what to extract
Pick a document type or write a free-text instruction. Custom mode shows you the inferred schema before extraction.
Step 3
Export the result
JSON, CSV, or Excel — whichever fits where the data is going.

Common use cases

Receipt photos — line items, totals, payment method

Photos of business cards — name, role, company, contact

Photos of IDs and passports — identity fields and dates

Screenshots of tables — turn dashboard data into rows

Photos of menus — items, prices, descriptions

Photos of forms and intake sheets

Whiteboard photos — capture handwritten notes as text

Product label photos — ingredients, nutrition facts, expiry

Photos of charts and graphs — extract the underlying values

Historical document scans — letters, ledgers, certificates

Sample output

Example: extracting a table from a screenshot

Request: "extract this table as rows"

Result:
{
  "rows": [
    { "country": "United States", "users": 18420, "revenue_usd": 142800 },
    { "country": "United Kingdom", "users": 6310, "revenue_usd": 51200 },
    { "country": "Germany", "users": 4980, "revenue_usd": 38900 },
    { "country": "Australia", "users": 2150, "revenue_usd": 17400 }
  ]
}

Frequently asked questions

How do I extract data from an image?+

Upload the image (PNG, JPG, WEBP, HEIC), describe what you want extracted, and download the result as JSON or Excel.

Does it work on phone photos with glare or at an angle?+

Yes. The model is trained on real-world photos, not just clean scans. Sharper photos give cleaner extractions.

Can it extract a table from a screenshot?+

Yes. Screenshots of tables — from a webpage, a dashboard, or a PDF viewer — extract into rows and columns just like a real spreadsheet.

Does it read handwriting?+

Yes — printed handwriting in English and most Latin-script languages works well. Cursive and faded handwriting are harder. For critical handwritten documents, review the output before relying on it.

What image formats are supported?+

PNG, JPG, WEBP, and HEIC (the format iPhones save photos in by default). Up to 20 MB per image. For multi-page scans, save as PDF and use the PDF extractor instead.

Can I extract data from multiple images at once?+

The web app processes one image at a time. The REST API on the paid plan handles bulk image extraction with a stable schema across calls.

How do I extract just the text from an image?+

Use the image-to-text extractor — same Gemini engine but tuned for plain-text output (with handwriting, multi-language, and translation modes). For structured fields, stay on this page.

How do I extract just the numbers from an image?+

The image-to-text extractor has a numbers-only mode that returns just the digits along with their nearby labels — useful for meter reads, totals, scoreboards, and any image where the words around the numbers don't matter.

How do I extract data from a chart or graph image?+

Use the chart data extractor — it reads axis labels and series legends to return the underlying numbers as a table. Bar, line, pie, and scatter charts all work.