Human transcripts of interviews, depositions, dictation and evidentiary audio — in 326 languages.
Last reviewed August 23, 2026 by the Prism Linguistics editorial team
Audio transcription services turn a recording into an accurate written transcript. Prism Linguistics transcribes interviews, depositions, medical dictation, focus groups, podcasts and evidentiary audio in 326 languages, using human transcriptionists rather than software alone. Most files come back within 2 to 5 business days, and quotes arrive within 60 minutes during business hours.
A recording sitting on a drive is hard to work with. A transcript is something you can actually use: an attorney can cite it, a researcher can code it, a producer can pull quotes from it. Getting the style right for that purpose is half the job, and it's the first thing we'll ask you about.
Every project goes to a transcriptionist who works with that kind of audio regularly.
Research interviews, oral histories and focus groups, with consistent respondent labels and timestamps ready for qualitative coding software.
Depositions, hearings, arbitrations and attorney-client interviews, taken verbatim where the wording has to survive scrutiny. See our legal language services.
Physician dictation, case notes, IME reports and recorded consultations, handled with HIPAA-aware care. More on our healthcare language services.
Episode transcripts for show notes, accessibility and SEO, plus raw interview tape for editing. Related: media translation.
Board meetings, earnings calls, panel sessions and all-hands recordings, turned into a written record that stands up as minutes.
911 calls, body-worn camera audio, custodial interviews and wiretap material, handled with chain-of-custody care. See law enforcement services.
This is where we differ from a typical US transcription company. Foreign-language audio is a two-stage job: someone has to write down exactly what was said in the source language before anything can be translated. That first stage goes to a native speaker of the language on the recording, drawn from linguists working across 326 languages.
From there you have three options: a transcript in the original language only, an English translation only, or both side by side. Attorneys and research teams reporting to an IRB usually want both, because the original is the evidence and the translation is the interpretation. The translation stage runs through the same reviewed workflow as our document translation services, and where a court or agency needs it we attach a signed Certificate of Translation Accuracy, as used in our certified translation work.
A practical example: a Houston law firm holds a 40-minute recorded phone call in Spanish. We deliver the Spanish transcript with timestamps, an aligned English translation and a certification page, so the exhibit is ready for filing without a second vendor.
Picking the wrong style is the most common reason a transcript gets sent back, so we ask what it's for before we start.
| Feature | Verbatim | Clean Read |
|---|---|---|
| What it captures | Every word as spoken: false starts, repetitions, fillers ("um", "you know"), stammers, with non-verbal sounds noted | The speaker's own words and structure, with fillers, stumbles and repetitions removed |
| Best for | Depositions, 911 and body-cam audio, custodial interviews, disciplinary hearings, linguistic research | Business meetings, research interviews, podcasts, dictation, oral histories |
| Reads like | A court record: faithful but slow to read | Natural written speech: quotable and easy to skim |
| Relative cost | Higher, because it takes longer to produce | Standard |
A simple rule: if the transcript could end up in front of a judge or a regulator, choose verbatim. If it's going to be read, coded or quoted from, choose clean read. Tell us the purpose and we'll recommend one.
If it plays, we can usually transcribe it.
Turnaround depends on runtime, audio quality and style. Most standard files return within 2 to 5 business days. Same-day and next-day delivery are often possible for short, clear recordings, subject to capacity, so tell us your deadline up front and we'll confirm it in writing. For anything large we'll set up a secure upload link rather than email.
Two boundary cases worth knowing. If your source is footage rather than audio, our video transcription service works from the file directly and can timestamp against the picture. If the goal is text on screen for viewers rather than a document, that's subtitling and captions, a different craft with its own reading-speed rules.
Honest answer first: for a clear English recording of one person speaking into a decent microphone, automated transcription has become good and cheap, and an app may serve you fine. We'd rather say that than pretend otherwise.
Human transcription earns its cost where automated tools reliably fail: overlapping speech, heavy accents, background noise, code-switching between languages mid-sentence, drug names and case citations, and telling apart six voices in a conference room. It also matters when an error has consequences. An AI tool that confidently mishears a dosage in medical dictation, or a name in a 911 call, produces a transcript that looks clean and is wrong in a way nobody catches.
Our process puts two people on every file: a transcriptionist produces the draft and a reviewer checks it against the audio before delivery. Where a section genuinely can't be made out, we mark it inaudible with a timestamp rather than guess. A visible gap is honest; a confident guess is not.
Recordings are often more sensitive than documents: a voice identifies a person, and people say things aloud they would never write down. Access is limited to the transcriptionist and project manager on the job, transfer runs over secure links, and files are deleted from active systems after delivery.
For medical audio we work to HIPAA-aware practices and will sign a business associate agreement where your organization requires one. For legal and law-enforcement material, including body-cam and custodial interview audio, we apply chain-of-custody care: no copies beyond the working team, logged handling, and transcripts formatted so they can be exhibited. Non-disclosure agreements are routine on this work, not a special arrangement.
Four steps from recording to transcript.
Upload the file through the quote form or ask us for a secure link for large recordings. Tell us the runtime, the language, roughly how many speakers there are, and what the transcript is for.
We recommend verbatim or clean read, confirm timestamps and speaker labels, and reply with a written quote and delivery date, usually within 60 minutes during business hours.
A professional transcriptionist produces the transcript and a second reviewer checks it against the audio. Anything genuinely inaudible is flagged with a timestamp rather than guessed.
You receive the transcript in Word, PDF or your preferred format, with a translated version alongside the original if you ordered multilingual transcription.
As a guide, professional human transcription of clear English audio in the US typically runs about $1.50 to $3.50 per audio minute. Strict verbatim and legal formatting sit toward the top of that range or above it. Multilingual transcription is quoted as two stages, transcription plus translation, and usually lands higher depending on the language pair.
Four things move the number for any given file:
A short sample of your audio tells us more than any rate card. Your quote confirms the exact price before we start.
Tell us the runtime, the language and how many speakers. We'll reply with a written quote and a confirmed turnaround.
Get a Free Quote Call +1 (833) 282 8883