Transcribe user research interviews

Where this fits in a research workflow

Researchers usually want three things from a recording: a record of what was said, quotes they can stand behind, and a way to find a moment again without scrubbing. A transcript with a timestamp on every line gives you those. It does not tag themes, count how often something was said, or write the findings; that analysis stays yours, in whatever notes tool, sticky-note board or spreadsheet you already use.

From session to quotes

1. Record with the participant's agreement and tell them how the recording will be used. Download the audio or video file from your conferencing app; sessions are uploaded as files, not read from a link.
2. Upload it above. For a usability session with screen sharing, only the spoken audio is transcribed: what was on the screen is not read, so keep your own notes on clicks and screens.
3. With a free account, turn on speaker labels (up to eight) and rename them Moderator and P01, P02. Check the labels where voices overlap.
4. Read the draft with the audio on and correct product names, feature names and jargon.
5. Copy the quotes you need with their timestamps. Export Markdown or text for your notes, or CSV if you want one row per line in a spreadsheet for affinity mapping.

Searching across sessions

With a free account your transcripts stay in a library and search works across all of them, so a question such as which participants said "export" is a search, not a re-read. See how to search your transcripts. Keep a naming habit (study, participant code, date) from the first session; it is what makes the library usable later.

Participant privacy and client confidentiality

Recording rules differ between countries and between US states: in some places everyone in the conversation has to agree to be recorded, in others one person's consent is enough. Check the rule that applies to you before you record. This page is not legal advice. Participants' voices identify them, and interviews often include names, employers and details they would not want repeated. Replace names with participant codes in the edited transcript before it travels anywhere, keep the key that links codes to people somewhere else, and delete the uploaded file when your protocol says to. If you work for a client, check your contract or NDA before you upload their sessions: the audio goes to our server. For material that must not leave your machine, self-hosting is the route. We make no claim that the hosted site meets any client or regulatory requirement.

Drafts you review

The text is a draft. It can smooth a hesitation that mattered, miss a negation, or mishear a product name, and any of those can change a finding. No accuracy figure is offered. Before a quote goes into a report or a slide, play that second of audio. For academic fieldwork with an ethics board, read the research interviews page.

Which plan fits

Without an account you can try one file at a time, up to 15 minutes and 25 MB, 3 tries a day, with nothing to sign up for. Pro is $9.99 a month for 600 transcription minutes a month (10 hours of audio), with files up to 500 MB and 120 minutes each. Pro+ is $19.99 a month for 2,500 minutes a month, files up to 1 GB and 180 minutes each, and priority processing when the queue is busy. Paid plans are billed monthly in US dollars with PayPal and can be cancelled from your Billing page; transcripts you already made are not deleted when a plan ends.

Eight 45-minute sessions are 360 minutes, which fits inside Pro's 600 minutes a month; a 90-minute session fits inside a Pro file (120 minutes) but not inside the no-account preview.