Global Araç
Image To Text
Henüz görsel yok
OCR runs entirely in your browser via Tesseract WebAssembly. First run downloads the language pack (~10MB); later runs are instant.
Drop an image of text — a receipt, a screenshot of a chat, a scanned page, a photo of a sign — and the tool extracts the text using in-browser OCR (Optical Character Recognition). Supports 15+ languages: English, Spanish, French, German, Italian, Portuguese, Dutch, Russian, Polish, Turkish, Arabic, Chinese (simplified and traditional), Japanese, and Korean. Pick the language before scanning — wrong language guess yields garbled output.
Common uses: digitizing receipts and invoices for expense tracking, extracting quotes from screenshots of articles or social media, copying text from images in PDFs that aren't OCR-searchable, capturing a recipe from a magazine photo, or reading a sign in another language (then paste the result into a translation tool).
The OCR engine runs entirely in your browser via WebAssembly (Tesseract.js, the JavaScript port of the Tesseract OCR library). The first scan downloads a ~10 MB language model and takes a few seconds; subsequent scans in the same language reuse the cached model and run in 1–3 seconds for a typical screenshot. No image leaves your device — useful for receipts and documents you don't want to send to a cloud service.
Nasıl Kullanılır
- Pick the language of the text in the image. Wrong language picks produce garbled output, so this matters.
- Drop the image into the upload area or click to browse. Best results: clear focus, good contrast, ~300 DPI scanned-page resolution or sharper.
- Wait for OCR. First scan in a language downloads a ~10 MB model (a few seconds); later scans are 1–3 seconds.
- The extracted text appears in the output box. Edit it directly to fix any recognition errors before copying.
- Click Copy to put the text on your clipboard. For multi-column layouts, use Copy as Markdown to preserve some structure.
Ne Zaman Kullanılır
- Digitizing receipts, invoices, or business cards for expense tracking or contact import.
- Extracting text from screenshots of articles, social posts, or chat messages.
- Capturing text from images embedded in non-searchable PDFs (image-only scans).
- Quickly copying ingredients from a magazine photo or a recipe screenshot.
Ne Zaman Kullanılmaz
- Heavily stylized text (curved logos, decorative typography) — Tesseract is tuned for standard typography. Use a vision-LLM tool for those.
- Handwritten text — that's a different OCR mode; use the dedicated handwriting-to-text tool which is configured for it.
- Tables and forms where layout matters — output is plain text. For structured extraction (forms, invoices with positional data), use a specialized service like AWS Textract.
- Very low-resolution images (under 100 DPI equivalent) — accuracy drops sharply on blurry input.
Yaygın Kullanım Senaryoları
- Onboarding a colleague who needs the same calculation/conversion
- Verifying a number or output before passing it on
- Quick conversion during a typical workday
- Pre-decision sanity-check on inputs and outputs
Sık Sorulan Sorular
Why does the first scan take longer than the second?
The first scan in a given language downloads the OCR model (~10 MB compressed) into your browser cache. Subsequent scans in the same language reuse the cached model and skip the download — you go from ~5 seconds first time to 1–3 seconds repeat.
Will my image be uploaded to a server?
No. Tesseract.js runs entirely in your browser via WebAssembly. Open DevTools → Network during a scan and you'll see only the model download from a CDN — never your image. After the first scan, even the model is cached locally.
Why is accuracy lower for handwriting?
Tesseract is trained primarily on printed/typed text. Handwriting needs a different kind of training data and a different segmentation approach (cursive flow rather than discrete glyphs). Use the dedicated handwriting-to-text tool, which uses a model tuned for it.
What's the maximum image size?
Limited by browser memory. In practice ~10 megapixels works fine; above that, processing slows and may run out of memory on phones. If you have a giant scan, downsize to ~3000 px wide first — OCR accuracy doesn't increase past that resolution.
Why does the output have weird line breaks?
OCR sees visual lines, not logical paragraphs — so wrap-around words appear on separate lines. The Copy as Markdown option does best-effort paragraph reflow. Or just paste into a text editor and join the broken lines manually.
Can it handle multiple languages in one image?
Pick the dominant language for best results. Mixed-language images (e.g. an English caption on a French photo) often do okay if the dominant script is the same (Latin), but accuracy drops on mixed scripts (Latin + Chinese) — those need a multi-language Tesseract config that's not exposed in this tool.