AI Utilities
AI-powered tools for speech and images. Speech Transcription runs entirely on-device, using browser-compatible models—your audio is never uploaded to our servers. OCR runs server-side for speed and accuracy, so it needs a network connection — see its page for what that means for your data.
OCR - Text Extractor
Extract text from images using server-side Tesseract OCR (requires a network connection). Upload JPG, PNG, or WebP images and get editable text back in seconds.
Speech Transcription
Convert speech to text in real-time using browser-based speech recognition. Works completely offline.
Why Use These AI Tools?
Traditional AI services require uploading your data to cloud servers, raising privacy concerns and requiring constant internet connectivity. Speech Transcription runs entirely in your browser using the Web Speech API, bringing AI capabilities directly to your device without uploading your audio. OCR runs server-side for better accuracy, with images handled ephemerally and never stored.
Benefits of Browser-Based AI:
- Complete Privacy for Speech Transcription: Your audio is processed locally and never uploaded to our servers
- Offline Capability: Speech Transcription works without internet after initial page load; OCR needs a connection, since it runs server-side
- No API Costs: Free to use, with generous rate limits
- Mobile-Friendly: Both tools work well on mobile browsers
Technologies We Use
Native Tesseract for OCR
OCR runs native Tesseract on our server rather than a browser-based WebAssembly port — generally faster and more accurate, especially on lower-end or mobile devices. Your image is handled ephemerally — processed in memory and never stored; see our Privacy Policy for details.
Web Speech API for Transcription
The Web Speech API is built into modern browsers, providing real-time speech recognition without any downloads. It supports multiple languages and works on both desktop and mobile.
Frequently Asked Questions
Is my data really private?
For Speech Transcription, yes: all processing happens entirely in your browser using the Web Speech API, so your audio is never uploaded to our servers. OCR is different — it uploads your image to our server to run native Tesseract, handled ephemerally: held in memory only for the duration of that request, never written to disk or a database, and discarded once the extracted text is returned. Either way, we don't store or share your personal information — like most websites, our server logs record visitor IP addresses for basic security and abuse prevention. See our Privacy Policy for details.
Do these tools work offline?
Speech Transcription does — once the page loads and the browser's speech model is ready, you can use it completely offline. OCR does not: it needs a network connection, since it runs on our server rather than in your browser.
How accurate is the OCR?
Native Tesseract provides excellent accuracy for printed text, typically 95%+ for clear images. Accuracy depends on image quality, font clarity, and language. For best results, use high-resolution images with good contrast.
Will this work on my mobile phone?
Yes! Both tools work well on mobile browsers — OCR especially, since running Tesseract on our server avoids the slower processing times in-browser OCR can have on lower-end devices.
What image formats are supported for OCR?
The OCR tool supports JPG, PNG, WebP, and other common image formats, up to 20MB. For best results, use images with clear, high-contrast text.
Which browsers support speech transcription?
The Web Speech API is supported in Chrome, Edge, Safari, and most modern browsers. You'll need to grant microphone permission for the transcription to work.