Frequently asked questions

A service that turns a court transcript (PDF or text) into read-along audio, with a different voice for each speaker, plus a private page where you can click a line or word to jump to it.

Uploads are invite-only for now. Email [email protected] to ask for an invite code.

PDF or plain text, up to 40 MB. Scanned PDFs are read with OCR, and if a scan is too poor to read reliably we stop and tell you instead of guessing.

Maximum file size is 40 MB. That's about 1,000 pages of a typed transcript, or about 40 to 130 pages if your document is a scan (a PDF made from photos of each page). Not sure if yours is a scan? Try selecting a word with your mouse. If you can't highlight any text, it's a scan.

The price depends on how many characters will be spoken. You see it before you agree, and nothing is charged until you press Make the recording. Launch pricing is $0.16 per 1,000 characters with a $10 minimum per document. See the Pricing page.

The price usually appears within a short while. Making a short recording takes a few minutes. The page shows progress and a rough estimate, and you do not have to keep it open. Longer documents take longer.

We email a private link. Check your junk or spam folder if you do not see it.

Anyone who has your private link, until the files are deleted. Nobody browses the files and there is no public list. Do not forward the link unless you mean to share it.

Automatically 48 hours after the recording is ready, or immediately when you press Delete this audio now. An upload you never confirm is deleted within 24 hours.

Yes. Press Download MP3 on your listen page. It is saved under your uploaded file's name, for example VV test hearing.txt becomes VV test hearing.mp3. Download before the 48 hours are up.

No. Verbatim Voices does not train on your documents. The text is sent to Amazon Polly only to make the voices. See the Privacy page for details.

No. It is a convenience copy read by computer voices. The written transcript is the official record, and this is not legal advice.

Mostly, but not always. Voices read what the text says, and unusual names, citations and numbers can be mispronounced. A scan can contain OCR mistakes. Have a person check anything you will rely on. See Accuracy and limits.

Amazon Polly today. We may offer other voice engines later, and if we do you will see the price before you agree.

Not yet. Highlighted video, SRT and VTT captions, a full download package and an accuracy report (Places the listener still wants a person to check) are planned and not available now.

Write to [email protected] with the time and what you saw. Please do not send the document itself unless we ask.