Screenshot to Document
Free OCR powered by Tesseract — runs entirely in your browser, nothing is uploaded to any server
Drop image here or browse files
PNG · JPG · WEBP · BMP · TIFF · Max 20 MB
What does this tool do?
This tool uses Tesseract.js — a free, open-source OCR (Optical Character Recognition) engine — to read text from your image. It then applies smart parsing to detect structure: rows, columns, line items, and headers. The result is exported as a clean spreadsheet, JSON object, or plain text.
Everything runs inside your browser. No image is ever sent to a server. No API key or account is needed — ever.
Supported document types
| Type | Examples | Best output |
|---|---|---|
| 📊 Table | Excel screenshots, HTML tables, data grids | CSV / Excel |
| 🧾 Receipt | Shop receipts, restaurant bills, petrol slips | CSV with Item/Price columns |
| 📄 Invoice | Freelance invoices, utility bills, tax invoices | JSON with line items + totals |
| 🎓 Marks sheet | Report cards, exam results, grade tables | CSV with Student/Subject/Marks |
| 🗓️ Timetable | Class schedules, rosters, weekly planners | Table preview / CSV |
| ✍️ Notes / list | Bullet lists, to-do items, numbered lists | Plain text / JSON array |
How it works
Upload your image
Select any screenshot or photo. The image stays entirely in your browser — it is never uploaded to any server.
OCR reads the text
Tesseract.js loads in the browser and scans the image, detecting every character with position and confidence data.
Parser detects structure
The raw OCR output is analysed for rows, columns, separators, and patterns — turning a flat list of words into a structured table.
Download your format
Switch between Table, CSV, JSON, and plain text views. Download the format you need or copy to clipboard.
Tips for best results
- Screenshots work better than photos — screen captures are perfectly sharp; phone photos of screens can have glare and distortion.
- Crop tightly — remove headers, footers, and unrelated screen areas before uploading.
- Higher resolution = better accuracy — use 1×/2× screenshots; avoid heavily compressed JPEGs.
- Select the correct document type — this tells the parser how to interpret the lines it finds.
- Dark text on white background — OCR accuracy is highest with high-contrast, printed text.
- For handwritten notes — print clearly; cursive is harder for OCR engines to read reliably.
Frequently asked questions
Yes. This tool uses Tesseract.js, which is a free open-source OCR library that runs in your browser. There is no API key, no account, no usage limit, and no paid tier. It costs you nothing to use it as many times as you like.
No. The image is read by JavaScript running inside your browser tab. It never leaves your device. ToolBase's servers never see your image or its contents.
On the first extraction, Tesseract.js downloads its language model (about 10 MB). This is cached by your browser after the first use, so subsequent extractions are much faster.
Download the CSV file, then in Excel go to File → Open and select the .csv file. Excel will detect the comma-separated structure automatically. You can also drag the CSV file directly onto an open Excel window.
By default the tool uses the English OCR model, which works well for most Latin-script languages. Tesseract supports 100+ languages — if you need Arabic, Urdu, Chinese, etc., contact us and we can explore adding a language selector.
OCR accuracy depends on image quality. Blurry, low-contrast, or stylised fonts reduce accuracy. The confidence score shown after extraction tells you how confident the OCR engine was. For low-confidence results, try a sharper or higher-resolution image.
What makes this tool different
Most image-to-text tools simply dump extracted characters into a single block of unformatted text. This tool goes further: it analyses the spatial arrangement of recognised text in your image — identifying which words share the same horizontal line (rows) and which are vertically aligned (columns) — then reconstructs that structure as a proper table you can open in Excel or import into any database.
This means a photograph of a marks sheet comes back as a spreadsheet with student names in one column, subjects across the top, and scores in the cells — not a wall of scrambled text you'd need to reformat manually.
Supported document types and best results
| What you have | Best image source | Output format |
|---|---|---|
| Data table (Excel screenshot, website grid) | Desktop screenshot | CSV — opens directly in Excel/Sheets |
| Receipt or till slip | Phone photo, well-lit | Plain text or CSV with line items |
| Invoice or purchase order | Scan or phone photo | CSV with vendor, date, amounts |
| Marks sheet or result card | Phone photo or scan | CSV with student, subject, score columns |
| Timetable or schedule | Screenshot or photo | Plain text or structured CSV |
| Handwritten list or notes | Phone photo, good light | Plain text (print clearly for best accuracy) |
| Business card | Phone photo | Plain text with name, phone, email |
| Locked PDF page | Screenshot of the page | Plain text or CSV depending on content |
Your data never leaves your device
Unlike most OCR tools that upload your image to a cloud server for processing, this tool runs the entire extraction process inside your browser using a JavaScript OCR library. Your image is never uploaded to any server — not ToolBase's, not anyone else's.
This makes it safe to use for sensitive documents: payslips, bank statements, medical reports, confidential invoices, and private correspondence. Once you close the browser tab, nothing is retained anywhere.
Getting accurate results — the most important tips
Use screenshots, not photos of screens
This is the single biggest accuracy improvement available to most users. A screenshot is pixel-perfect — every character is sharp, there's no reflection, no parallax, no angle distortion. A photo of a screen introduces moiré interference patterns, glare from the screen's backlight, and geometric distortion from the viewing angle. Whenever the data exists on a screen, take a screenshot rather than photographing the screen.
How to take a screenshot: Windows (Win+Shift+S), Mac (Cmd+Shift+4), iPhone (Side button + Volume Up), Android (Power + Volume Down).
Crop tightly before uploading
Remove everything except the data you need. Navigation bars, page headers, surrounding whitespace, and unrelated content all give the OCR engine more to process and more opportunities to misinterpret layout. A tightly cropped image focused on the table or text produces significantly cleaner output.
Check for these common OCR errors
Even with a perfect image, certain characters are frequently confused. Always scan your extracted data for:
- 0 vs O — zero and the letter O look identical in many fonts
- 1 vs l vs I — one, lowercase L, and uppercase i are often indistinguishable
- 5 vs S and 8 vs B — in compressed or stylised fonts
- Decimal points and commas — particularly in numbers like 1,234.56 where the separators may be misread
For financial data — receipts, invoices, marks totals — always verify that the extracted numbers sum correctly before using them in a spreadsheet or report.
Practical workflows
WhatsApp screenshot → spreadsheet
Someone sends a price list, schedule, or result table as a WhatsApp image. Open it on your phone or desktop, take a screenshot (or save the image), upload it here, and download as CSV. Open the CSV in Excel or Google Sheets. What would take 20 minutes of manual typing takes under 60 seconds.
Paper receipt → expense report
Photograph the receipt in good light with your phone camera parallel to the paper. Upload the photo, select Receipt mode, and extract the vendor name, date, and line items. Copy the amounts directly into your expense tracking spreadsheet without retyping.
Scanned exam result sheet → marks database
A school result sheet scanned as a PDF or image can be uploaded directly. The tool extracts student names, subjects, and scores into a structured CSV that you can open in Excel and immediately sort, filter, calculate averages, and generate charts from — without a single number typed by hand.
Locked PDF → editable text
Some PDF files have text selection and copying disabled. Take a screenshot of the relevant page (or pages), upload it here, and extract the text. The output won't preserve the original formatting but gives you editable, searchable text immediately.
Output formats explained
CSV (Comma-Separated Values)
The most useful format for tabular data. Open it by double-clicking in File Explorer (Windows) or Finder (Mac) — Excel opens it automatically. In Google Sheets, go to File → Import → Upload → select your CSV file. Each row in the image becomes a row in the spreadsheet; each column becomes a column.
JSON
Best for developers. If you need to process the extracted data in a Python script, a web application, or a database import, JSON gives you a structured array you can parse directly. Each row is an object with keys matching the detected column headers.
Plain text
Formats the data as a monospaced ASCII table — rows separated by line breaks, columns aligned with spaces. Best for pasting into a document, email, or chat message where you want the table to be readable without needing to open a spreadsheet.