Questions, answered plainly

A Word document. You upload a recording and one finished transcript from the client whose format you need, and about a quarter of the recording's length later you download a draft laid out in that format: their caption page, their speaker labels, their indenting, their certificate page. With it comes a plain-text file of review notes listing every word the transcription was unsure of, with a timestamp, so your proofer listens to those spots first. The first job is free, so the quickest way to see one is to run your own.

It's free, there's no card, and you don't join anything. You sign up, we approve the account (usually within the hour), you upload one real job, and the draft is yours whether or not you ever come back. The first job is capped at 90 minutes of audio; if yours is longer we'll say so before processing and you can trim it or add a card for the remainder.

No, and it isn't meant to. It's a first draft. You review it against the audio, correct it, and certify it exactly as you do today. What changes is that you start from a document instead of a blank page. The transcriber who has timed her own jobs reports her review running a quarter to two-thirds shorter than typing them. That's one person's experience, and yours may differ, which is why the first job is free.

Any common audio or video: MP3, WAV, M4A, WMA, MP4, MOV and others. Courtroom FTR (.trm) recordings upload directly as the whole set of files, no export step; multi-microphone days are handled as such. You can also upload several recordings, or a zip of them, as one job: they're stitched in order and the draft gets a heading where each recording begins. Up to 3 GB per upload.

Usually a quarter of the recording's length: a one-hour interview is typically back in fifteen to twenty minutes, a five-hour court day in about an hour. You get an email when it's ready. Within 24 hours is our standard commitment for the rare case where it takes longer.

No. The draft is a .docx. Review it in Word alongside ExpressScribe or whatever player you already use. Nothing to install, no new editor to learn.

On clear interview audio the draft's words match a transcript typed from scratch around 97 to 98 percent of the time. On phone lines, cross-talk and echoing rooms it's lower, and those are exactly the spots the review notes flag. We'd rather you judge it on your own audio than on our number, which is what the free job is for.

No. A technical guard means the AI may only adjust punctuation and capitalisation. Every spoken word, every false start and repetition, survives in the order it was said. If the client's house style is "clean verbatim", fillers are removed and nothing else. Testimony is never paraphrased.

From one finished transcript you upload with the job, Word or PDF. It reads the layout: title page, how speakers are labelled (Q./A., NAME:, THE COURT), the indenting and spacing of the body, the certificate page, and reproduces them. If your template is a blank shell with no sample dialogue, it recognises a court caption from the cover page and uses court conventions. Long speeches are broken into paragraphs of roughly a hundred words.

The draft does, from context: who asks and who answers, names when they're spoken, the judge, clerk and counsel in a hearing. It uses the label style your template uses. When a voice can't be placed it's labelled UNIDENTIFIED SPEAKER rather than guessed. Names are the part most likely to need your eye; roles are usually right.

A text file listing each word the transcription was least sure of, with its timestamp and the turn it sits in. Optionally the same words are highlighted in the draft. It turns "listen to the whole thing" into "listen here, here and here".

Tell us, with the thirty-second feedback form beside the draft or by replying to the delivery email. If it was our fault, we re-run it at no charge. Several of the things this product does today exist because a transcriber told us a draft was more trouble than typing it.

By audio minute: the length of the recording you upload, rounded to the nearest minute. A 36-minute interview is 36 minutes whether it takes us eight minutes or twelve to process.

On your billing anniversary, not the calendar month, so the counter in your account always matches your invoice.

Nothing stops. The extra minutes are billed at your plan's per-minute rate and appear on your next invoice. There is no hard cap and no "upgrade to continue" wall.

Any time, from your billing page. Upgrades and downgrades are prorated by Stripe and take effect immediately. If you cancel, you keep access until the end of the period you've paid for. A founding rate stays with you for as long as you stay subscribed, including if you move between plans.

Not for the free first job. You add a card when you choose a plan, and it's handled by Stripe; we never see your card number.

If a draft was unusable because of a fault on our side, we re-run it free or credit the minutes. Subscriptions can be cancelled at any time and aren't renewed after that. If something else has gone wrong, write to us; we'd rather fix it than argue about it.

Onto servers we operate in the United States, and nowhere else. Transcription, speaker identification and formatting all run on our own hardware. Your recordings are never sent to OpenAI, Google or any third-party AI service. Details on the Security page.

No. Your audio and documents produce your transcript and nothing else. There is no model training and no "service improvement" fine print.

Audio, template, draft and notes are deleted from our active systems 14 days after delivery, automatically. You can delete a job's files yourself sooner from your dashboard. Encrypted backups expire within 30 days.

You do, and your client does. We hold no rights in your recordings or the drafts we produce from them.

No. We are a small company and we'd rather say that than borrow a badge. What we do instead: no third-party AI, US-only storage, encryption in transit, automatic deletion, access limited to the people who operate the service, and a signed NDA you can put in place online in about a minute before you upload anything. If a client needs more, ask and we'll answer in writing.