UK Transcription Services: Human Transcripts of Audio and Video
Transcription services turn an audio or video recording into written text. Espresso transcribes interviews, meetings, focus groups, medical dictation, research recordings and broadcast material for UK-based clients in 150+ languages, with a professional transcriber typing and checking the file rather than speech recognition alone. Transcription starts at £5 per audio minute, with the exact quote set by the audio quality, the number of speakers, the subject matter and the turnaround, and your itemised quote reaches you within 1 hour. If the transcript then has to be translated, standard translation is £0.10 per word and a certified translation is £30 per page of up to 250 words, delivered in 24–48 hours with a second linguist checking it under ISO 17100.
ISO 17100 Certified
2,000+ Linguists
NDA on Every Job
UK GDPR Handling
150+ Languages
Quote in 1 Hour
Request a Transcription Quote Now
We’ll get back to you within 1 hour
What is a transcription service? A transcription service converts an audio or video recording into written text, either word for word or tidied into a readable transcript. Espresso uses human transcribers rather than automatic speech recognition, has a second linguist check the file under ISO 17100, covers 150+ languages, and handles every recording under UK GDPR. Professional transcription services pair a typist with a checker, which is how you get accurate transcripts instead of a raw machine draft. Transcription starts at £5 per audio minute and a free itemised quote arrives within 1 hour.
Why Do UK Organisations Send Recordings to a Human Transcriber?
Because in the UK a recording is usually evidence, research data or a clinical record before it is a document, and each of those carries rules about who may hear it and how the audio file is stored. Espresso gives every job to a professional transcriber who writes British English, knows the industry-specific terminology of the field and has signed a confidentiality agreement before the audio moves. Nothing is left as raw machine output.
Our transcription team works on witness and client interviews, board meetings, disciplinary and tribunal recordings, interviews and focus groups, lectures, clinical dictation and broadcast footage. Every transcript is read back against the audio by a second person before delivery, the same two-pass discipline that ISO 17100 sets for translation, and you choose the format: speaker labels, timestamps at a set interval, or a clean reading copy.
Data security is part of the work rather than an extra. Each project runs under a signed non-disclosure agreement, files move by encrypted transfer, access is limited to the transcriptionists and project manager on your job, and audio is returned or deleted on your instruction. Recordings are never fed to speech recognition training sets, which matters when the audio contains a patient, a suspect or a research participant.
A transcript often has to exist in a second language, and we do both halves in one project. A Polish witness interview can be transcribed in Polish and translated into English for a UK court bundle; a Japanese focus group can come back as a transcript plus an English report. One project manager, one NDA and one invoice cover the whole job, and the transcriber’s notes on unclear audio reach the translator instead of being lost between two suppliers.
Our UK transcription work covers 150+ languages, drawing on more than 2,000 linguists. The human transcription services UK clients order most are same-language transcripts and transcription combined with translation, and both run as one project rather than two. Most projects come back in 2–3 working days, a quick turnaround is possible on shorter files, and a 24-hour express option is available where the audio allows it. Clients that are UK based get a project manager working their hours, so speed and accuracy are settled before work starts rather than argued about afterwards.
Ask us to transcribe your audio or video file and an itemised quote reaches you within 1 hour.
Which UK Sectors Do We Transcribe For?
Each sector has its own vocabulary, layout conventions and rules about confidentiality, so every recording goes to a transcriptionist who already works in that field. The five below account for most of what UK clients send us.
What Do UK Courts, Tribunals and Police Recordings Require?
A transcript of a court or tribunal hearing in England and Wales is ordered from HM Courts and Tribunals Service on form EX107, and only transcription companies contracted to the Ministry of Justice may produce one from the official recording; a judge decides requests that involve a private hearing, and parties are not normally allowed to record a hearing themselves. Espresso is not on that framework and does not transcribe HMCTS recordings. What we do transcribe is the audio a UK legal team holds lawfully: the copy of a police interview supplied to the defence, which the Home Office code of practice on audio recording interviews with suspects, PACE Code E, requires the police to seal as an exhibit and to release once someone is charged; arbitration and mediation sessions recorded with the parties’ agreement; insurance and loss-adjuster interviews; internal investigations; and client meetings recorded with consent.
How Is Your Recording Handled Under UK GDPR?
A recording of someone who can be identified is personal data under UK GDPR, and if it covers health, criminal allegations or similar ground the Information Commissioner’s Office treats it as special category data needing a separate condition to process. When you send us audio we act as your processor, which the ICO says must be set out in a written contract, so the terms covering purpose, security, sub-processors, retention and deletion are agreed before the first file moves. UK GDPR is a law we work to, not a certificate anyone holds.
In practice that means a signed NDA on every job, encrypted upload and download, access restricted to the named transcriptionists and the project manager, and audio and transcripts deleted or returned when you say so. Sub-processing is disclosed rather than assumed. If your organisation runs its own security assessment for suppliers handling patient or client recordings, ask at quote stage and we will complete it.
University research adds a second layer. UKRI’s Economic and Social Research Council expects researchers to explain to participants what the limits of confidentiality are, to say when material will be anonymised, and to cover confidentiality in the data management plan, and recordings of identifiable voices are among the things that can trigger a full ethics review. We sign the confidentiality agreement your ethics committee asks for, transcribe research interviews under it, pseudonymise names where the protocol says so, and delete the audio on the date you set.
Which Type of Transcript Should You Order?
The right transcript type depends on what the text is for: evidence, analysis, publication or subtitles. All four types below can be produced from the same audio recording, in any of the types of audio and video files we accept, and you can mix them across a digital transcription programme.
Verbatim Transcription
Verbatim transcription records every word as spoken, including repetitions, false starts, filler words, hesitations and audible reactions, with non-speech sounds marked. It is the format for evidence, regulatory investigations and linguistic or qualitative analysis, where how something was said matters as much as what was said.
Intelligent Transcription
Intelligent transcription keeps every point the speaker made but drops the stumbles, repetitions and filler words that carry no meaning. It is the usual choice for meetings, interviews, webinars and research summaries, and it is the fastest of the formats to read.
Edited Transcription
Edited transcription goes a step further and turns speech into prose: sentences are tidied, grammar is corrected and asides are dropped, without changing what the speaker meant. Use it for material that will be published, circulated to a board, or used as the source text for a translation.
Timestamped and Speaker-Identified Transcripts
Timestamps can be inserted at a fixed interval or at every speaker change, and each voice can be labelled by name or by role. Tribunal bundles, focus groups and panel discussions are much easier to navigate this way, and a timecoded transcript is what an editor needs to cut video or build captions. Both options can be added to a verbatim, intelligent or edited transcript.
What Does Transcription Cost in the UK, and How Fast Is It?
Transcription starts at £5 per audio minute. Rates in the UK depend on audio quality, the language, the subject, how many people are speaking, the turnaround time and the formatting you need, so where a recording sits above that floor is set by the file itself. A tidy dictation costs less per minute than a panel discussion. Send the recording and you will have an itemised quote within 1 hour, with the fee agreed before any work starts.
Four things move a quote most: background noise and poor audio quality, multiple speakers talking over each other, specialist legal or medical terminology, and detailed timestamping. Clear audio from a one-to-one interview costs less to transcribe than the same minutes of audio from a noisy six-person meeting, which is why we listen to a sample before quoting. A fast turnaround is quoted on the file rather than added as a standing surcharge.
Most transcripts are delivered in 2–3 working days, with a 24-hour express option for shorter recordings received early in the day; a faster turnaround depends on the length and the state of the audio. If the transcript is then translated, standard translation is £0.10 per word, a certified translation is £30 per page of up to 250 words and comes back in 24–48 hours, and a sworn translation for a foreign authority is £55 per page and takes 7–14 working days.
For low-risk material such as internal webinars, an economy route is available: an automated transcription service produces the first draft and a native-language transcriber corrects it against the audio. It is not the right choice for evidence, clinical records or anything a regulator may read, and your project manager will say so rather than sell it to you.
150+
languages transcribed, with the same job able to carry transcription and translation under one project manager
1 hour
free itemised quote — the fee, the turnaround time and the delivery format confirmed before you commit
How Does Our Transcription Service Work, Step by Step?
Four steps: send the files, accept the quote, let a transcriber and a checker do the work, and receive the transcript in the format you asked for. Your project manager stays the same person throughout.
The four steps in full:
Step 1: Upload Audio or Video Files
Upload audio or video files through our secure quote form. We accept the common audio and video file types, including MP3, WAV, M4A, MP4 and MOV, plus exports from Zoom, Teams, Google Meet, dictation apps and conference systems. Tell us the accents, the number of speakers and any names or technical terms that appear, as that is what saves time later.
Step 2: Your Quote Within 1 Hour
A project manager listens to a sample, checks length, language, audio quality and formatting, and picks the transcriptionist with the right subject background. You receive an itemised quote within 1 hour, covering the fee, the turnaround time and the delivery format.
Step 3: Human Transcription, Then a Second Check
An experienced human transcriber produces the transcript in your chosen format, marking anything genuinely inaudible with a timestamp rather than guessing. A second linguist then checks the file against the audio for terminology, speaker labels and completeness before it is released.
Step 4: Your Transcript in the Format You Asked For
Transcripts are delivered as Word, PDF, TXT or SRT, in your house layout if you have one. Where the project includes translation, the original-language transcript and the translated version arrive together from the same project manager.
Is Human Transcription More Accurate Than AI?
For anything that carries consequences, yes. AI transcription tools and automatic speech recognition now cope well with one clear voice in a quiet room, and they are useful for a rough search copy of a webinar. They still lose accuracy on regional accents, crosstalk, background noise, legal and medical terminology and unfamiliar names, and they cannot tell you when they are wrong. A human transcriber flags the doubtful passage, keeps the confidentiality terms you signed, and produces text a court, an ethics committee or a regulator will accept as a record. Human-powered transcription costs more than a machine draft, and for evidence it is the only route to a highly accurate human transcription: professional transcribers who know the subject check every name, date and figure, which is what accurate transcription of a real recording takes.
Can ChatGPT Transcribe Audio?
It can produce a rough draft from a recording, but it is not a substitute for a professional transcription service where the material is confidential, regulated or technical.
The practical problems are the same as with any automated transcription: no accountable checker, no signed confidentiality terms, inconsistent speaker identification, and quiet errors in names, dosages and figures. Uploading a patient interview or a witness account to a general AI tool also puts personal data somewhere your data protection terms probably do not cover.
Which Related Services Can We Add to a Transcription Job?
Most transcription jobs arrive with something attached: a translation, a set of subtitles, or a second language version of the finished video. All of the services below run through the same project management team and the same confidentiality terms.
Specialist Transcription Services
Legal Transcription
For witness statements, arbitration recordings, disciplinary hearings and solicitor dictation, see our Legal Transcription service.
Podcast Transcription
For podcasts, interviews and spoken-word recordings that need show notes or a searchable text, ask about Podcast Transcription.
Multilingual Transcription
For recordings in more than one language, Multilingual Transcription delivers the original transcript and the translation in one project.
Government Transcription
For public sector meetings, consultations and inquiry material, ask about Government Transcription and the handling terms that go with it.
Academic Transcription
For interviews, lectures and fieldwork recordings tied to an ethics approval, see our Academic Transcription services.
Start Your UK Transcription Project
Subtitle Translation
For caption and subtitle files that follow Ofcom’s guidance on accuracy, timing and readability, including SDH, use our Subtitle Translation service.
Video Translation
To take spoken video content into another language after transcription, see our Video Translation services.
2,000+
Linguists
ISO 17100
Certified quality standard
1 hour
Free quote
150+
Languages Supported
What Do UK Clients Ask About Transcription Services?
How much does transcription cost in the UK?
Transcription starts at £5 per audio minute. The exact quote depends on audio quality, language, subject, number of speakers, turnaround and timestamping, so a clean dictation sits at the floor and a noisy multi-speaker recording above it. Send the file and the itemised quote arrives within 1 hour.
Can you transcribe a court or tribunal hearing?
Not from the official recording. A transcript of a hearing in England and Wales is ordered from HM Courts and Tribunals Service on form EX107 and produced by a transcription company contracted to the Ministry of Justice, with a judge deciding requests that involve a private hearing. We transcribe the recordings your side holds lawfully: client and witness interviews, arbitration and mediation sessions, and disciplinary hearings.
Do you transcribe police interview recordings?
Yes, from the copy released to the defence. PACE Code E, the Home Office code of practice on audio recording interviews with suspects, requires the police to seal the master recording as an exhibit and to provide a copy once a person is charged or told they will be prosecuted. We transcribe that copy verbatim, with speaker labels and timestamps, under a signed NDA.
Who transcribes my recording?
A native speaker of the language on the recording, matched to its subject, so legal, medical, research and broadcast material each go to someone who knows the terminology. A second person checks the finished transcript against the audio. Transcripts are delivered in British English spelling and conventions whenever you ask for that, and where the transcript is then translated, a second linguist revises the translation under ISO 17100.
How do you handle UK GDPR when I send a recording?
A recording of an identifiable person is personal data, and health or criminal material is special category data that needs a separate condition to process. We act as your processor under a written contract, as the Information Commissioner’s Office requires, with a signed NDA, encrypted transfer, restricted access, disclosed sub-processing, and deletion or return of the audio on your instruction.
How long does a transcript take?
Most transcripts are delivered in 2–3 working days. A 24-hour express option is available for shorter recordings sent early in the day. Long, multi-speaker or specialist recordings take longer, and your project manager confirms the date in the quote before work starts.
Can you produce subtitles that follow Ofcom’s guidance?
Yes. Ofcom’s Code on Television Access Services and its accessibility best practice guidelines ask for subtitles that follow the speech as closely as possible, keep filler words only where they carry meaning, stay readable, and describe relevant non-speech sound. We prepare timecoded transcripts and subtitle files with those points in mind, in English or translated.
What is the difference between verbatim, intelligent and edited transcription?
Verbatim keeps every word, including repetitions, false starts and audible reactions, for evidence and qualitative analysis. Intelligent transcription drops the stumbles and filler words while keeping every point made. Edited transcription turns speech into publishable prose without changing the meaning.
Can you transcribe a recording in another language and translate it?
Yes, in one project. A native speaker transcribes the original, then a translator works from both the transcript and the audio, so unclear passages are resolved rather than guessed. Standard translation is £0.10 per word, and a certified translation is £30 per page of up to 250 words, delivered in 24–48 hours.
Is AI transcription good enough for legal or medical recordings?
Not on its own. Automatic speech recognition handles one clear voice well, but loses accuracy on accents, crosstalk, background noise, drug names and legal terminology, and it gives no signal when it is wrong. For evidence, clinical records and research data, a human transcriber and a second checker are what make the transcript usable.
What audio and video files can I upload?
Common formats including MP3, WAV, M4A, AAC, MP4, MOV and WMV, plus exports from Zoom, Microsoft Teams, Google Meet, dictation apps and conference systems. Upload audio or video through the secure quote form; if the file is very large, ask your project manager for a transfer link.
