Human transcription costs roughly £1–£2 per audio minute, so £60–£120 an hour of recording. That reflects real time: four to six hours of careful work per audio hour on decent recordings, and considerably more on poor ones.
What it costs
What transcription costs — and what you're actually buying
What good looks like
- Timestamps, so you can find the audio again
- Speaker labels where there is more than one voice
- Uncertain passages marked rather than guessed
- The raw audio kept, not discarded after transcription
What your book will actually cost
Four questions — genre, length, what you have already, what you want at the end — and we send a costed plan for your book.
What you are actually buying
Transcription turns recorded speech into text. For a book it matters more than for anything else, because when audio is unclear an automatic system does not leave a gap — it invents a plausible word and presents it with the same confidence as the rest.
Why it costs that
Human transcription is priced per audio minute and reflects real time: four to six hours of careful work per hour of audio on decent recordings, and considerably more on poor ones. What you are paying for is accurate proper nouns, reliable speaker attribution, sensible handling of crosstalk, and judgement about what a mumbled phrase actually was.
How to spend less
Prepare a name list before you record — every person, place and recurring term. It improves any transcript, human or automated, more than any other single step. Then use automated transcription for the clear single-speaker recordings and pay a human for the difficult files that matter. And analyse the archive before transcribing it: you will use a fraction of what you have.
What good looks like
- Timestamps, so you can find the audio again
- Speaker labels where there is more than one voice
- Uncertain passages marked rather than guessed
- The raw audio kept, not discarded after transcription
Questions to ask them
- What do you do with unclear audio?
- Is this human or automatic, and if automatic, is it checked?
- Do you handle accents and dialect?
Red flags
- No marking of uncertain passages
- Per-minute pricing that does not distinguish difficult audio
- Anything that returns clean text from bad audio without queries
Honestly
Heavy accents, several people talking over each other, poor recordings and dense technical terminology all still favour a human transcriptionist. Read any transcript against the audio before quoting from it.
What Your Book Pro does here
Your Book Pro is a British book-writing platform. You bring a book — an idea, a folder of notes, half a draft, or a finished manuscript written somewhere else — and it gives you the editorial process a publisher would run: a plan, structured feedback, a simulated beta read, a line edit, a proofread, a cover, a blurb and a print-ready file. You can type it, upload it or speak it. The tools sit in one place and work on the same manuscript, so the feedback compounds instead of arriving as twelve disconnected opinions.
These prices are here to be used, not to sell you something — brief a freelancer with them, sanity-check a quote, or work out what you actually need. Where Your Book Pro changes the arithmetic, the page says so; where it does not, it says that too.
Heavy accents, several people talking over each other, poor recordings and dense technical terminology all still favour a human transcriptionist. Read any transcript against the audio before quoting from it.
What it costs
| What you are buying | Typical range | Notes |
|---|---|---|
| Automated | £0.10–£0.50 per audio-minute | Typical automated service pricing |
| Human, clean verbatim | £60–£90 per audio-hour | UK market for single-speaker clear audio |
| Human, difficult audio or multiple speakers | £90–£150 per audio-hour | UK market |
CIEP suggested minimum rates effective 1 March 2026; Books.by; publishing.co.uk; The Ghostwriters Agency. Dated 9 September 2026.
Upload your audio
Get my costed planQuestions people ask
- How much does transcription cost per hour of audio?
- £60–£120 per audio hour for human transcription. Automated is a small fraction, with accuracy depending heavily on the recording.
- Is automated transcription good enough for a book?
- For clear audio, close — except for proper nouns, which need a prepared name list and a check. For difficult audio, no.
- Should I transcribe everything I've recorded?
- Under about ten hours, yes. Beyond that, work out what the book is first and transcribe what serves it.
- What is clean verbatim?
- Every word the speaker said, with filler, stutters and false starts removed, and grammar and word choice untouched. It is the right default for a book.