Image to Excel

Image to Excel — extract tables from screenshots and images

Manually retyping data from a table screenshot into Excel is slow and error-prone. This tool uses OCR (optical character recognition) with table-structure detection to convert an image of a table into an editable spreadsheet. Upload a screenshot, scan the table, edit any misrecognized cells directly in the browser, then download as CSV for Sheets/Excel or XLSX for Microsoft Excel.

How it works

1. Upload your image — drag and drop a screenshot of a table, or click to browse. PNG, JPG, and WebP are supported.
2. Select the language — choose the language of the text in your image. 13 languages are supported including English, Chinese, Japanese, and major European languages.
3. Click Scan Table — the tool downloads a small OCR language model (first use only, cached afterward) and runs recognition entirely in your browser.
4. Review and edit — the detected table appears as an editable grid. Click any cell to correct OCR errors before exporting.
5. Download — export as CSV (universal, opens in any spreadsheet) or XLSX (native Excel format). Or copy as TSV to paste directly into Excel or Google Sheets.

Table detection: how it works

The tool runs Tesseract.js OCR with word-level position data, then clusters words into rows and columns based on their X/Y coordinates. It detects column boundaries by analyzing horizontal gaps between words, and groups words into rows by vertical proximity. The result is a structured 2D table that you can edit before exporting.

This approach works best on clean, grid-like tables with clear spacing between columns. It handles typical spreadsheet screenshots, PDF table exports, and website table captures.

Tips for best results

• Clean, high-contrast tables work best — dark text on a light background with visible grid lines or column spacing.
• Avoid skewed images — straight, upright text yields much higher accuracy than rotated or tilted images.
• Higher resolution — screenshots at 2x or higher resolution improve OCR accuracy, especially for small text.
• Crop tightly — crop to just the table area before uploading. Extra content around the table confuses the column detector.
• Check the result — OCR is not perfect. Review each cell and fix any errors before downloading. Numbers and symbols are most prone to misrecognition.
• Right language matters — select the correct language for your table text. Mixing languages in one image reduces accuracy.

Supported formats and languages

Upload images in PNG, JPG, or WebP format. Supported OCR languages: English, Chinese Simplified, Spanish, French, German, Japanese, Korean, Portuguese, Italian, Russian, Arabic, Hindi, Dutch. The language model downloads once (~10–15 MB) and is cached in your browser.

Privacy and browser-side processing

All OCR processing happens locally in your browser using Tesseract.js compiled to WebAssembly. Your image is never uploaded to any server. This makes the tool safe for sensitive data: financial tables, internal reports, personal spreadsheets, and confidential business data.

Common use cases

• Extracting tables from PDF screenshots into editable Excel spreadsheets
• Converting website table screenshots into CSV data for analysis
• Capturing financial data from bank statement screenshots
• Digitizing printed tables from photos taken with a phone
• Extracting data from presentation slides and reports
• Moving data from legacy system screenshots into modern formats

Related tools

People who extract tables from images also often use Image to Text, Image to JPG, Crop Image, JSON Formatter