Launching today
Scribo
Private AI transcription, OCR and subtitle translation
2 followers
Private AI transcription, OCR and subtitle translation
2 followers
Transcribe audio and video to text and timedsrt subtitles, pull text out of images and scanned PDFs, and translate text or subtitle files (English with Spanish, French, German). Whisper, Tesseract and Opus-MT run in your browser, so nothing is uploaded. Free up to 10 min per file, no account. Pro is $7/month for long files and the more accurate model.
Hey Product Hunt! I’m Lucas, a solo maker from France.
I kept needing three small things: a transcript of a voice memo or a call, the text out of a screenshot or a scanned PDF, and subtitles in another language. Most online tools want you to upload the file to their server. For client calls or personal documents, I didn’t love that.
So Scribo runs the AI in your browser: Whisper (via transformers.js, on WebGPU when available, WebAssembly otherwise) for audio and video to text and .srt, Tesseract.js + pdf.js for images and scanned PDFs, and Opus-MT to translate subtitle files while keeping the timestamps. Once the model is cached, it even works offline.
Free: 10 minutes per audio file, 3 pages of OCR and 2,000 characters of translation at a time, no account. Pro is $7/month (or $49/year) for 2-hour files, whole PDFs and the more accurate Whisper model.
Honest limits: no speaker labels yet, phones are slower than laptops, and machine translation needs a proofread.
I’d love your feedback: which device or browser did it struggle on, and what would you want next (summaries, speaker labels, more languages)?