Transcribe recordings without uploading them.

Meetings, interviews, support calls: voices you cannot send out of the building, turned into text on your own device. Export captions or text, and let a local AI sum it up.

No account. Free up to 30 minutes; longer recordings and exporting several files at once with the Orihako Pass.

What it does

• TranscriptionJapanese, English and more; name the language for better results
• Captions and textSRT, VTT and plain text
• Translated captionsChrome's built-in AI translates the captions
• Cut and splitPick what to keep from the transcript; drop the rest
• SummaryThe gist of a long recording, from an AI on your device
• Audio from videoDrop a video and take the sound as WAV
• On to the next toolAsk Chat about the transcript

How to use it

1. Press the button above: the studio opens on "Transcribe and add captions". Drop your audio file.
2. Choose the spoken language and press Transcribe. The first time, the AI model is fetched into this browser; after that it runs on the spot.
3. Fix any misheard words on the spot.
4. Save as SRT, VTT or text. It stays under "Continue" in the bar, so you can hand it to Chat.

Questions

Is my audio sent anywhere?

No. Transcription and summaries run on a model fetched once into your browser. It works with the network off, and the "Network" chip in the bar shows it.

Which audio formats work?

MP3, WAV, M4A, OGG, WebM and anything your browser can play. Drop a video and you can take just the audio.

How good is it in English?

A multilingual transcription model with a "high accuracy" option; naming the language helps. Long recordings take time in proportion to your computer.

Noise removal and loudness?

Not yet. Cutting, splitting and merging, transcription and summaries are what it does; if it is asked for, it will be added the same way, inside the browser.

Orihako Stream (video too) · The other Orihako tools · How your files stay on your device