Upload the interview recording
Choose MP3, WAV or M4A audio, or an MP4 or MOV with a readable audio track. Free files can be up to 10 minutes; Pro files can be up to 60 minutes.
Turn a recorded interview into editable text. Check each quote against the recording, add speaker names manually and export the reviewed transcript.
Original audio files: Free up to 25 MB and 10 minutes; Pro up to 100 MB and 60 minutes. Videos up to 500 MB, with the same plan duration limits. Each prepared audio part is limited to 25 MB. The original video is not uploaded. Browser decoding support is required.
Free: 1 file without an account, up to 10 minutes. Sign in for 30 free minutes each month. Noise removal and WAV download are always free.
Daily transcription capacity: Free shares 1,500 minutes; Pro has a separate 1,500-minute pool, independent of the free pool. Both reset at 00:00 UTC. If your pool is full, try again after the reset; no task is queued for the next day.
Checking your plan…
Good things start with a word.
Your transcript appears here.
Upload a recording to get started.
Start with a recording you own or have permission to process. This tool creates a draft for review, not a certified or guaranteed word-for-word record.
Choose MP3, WAV or M4A audio, or an MP4 or MOV with a readable audio track. Free files can be up to 10 minutes; Pro files can be up to 60 minutes.
Choose Transcribe recording and wait for the draft. Only prepared audio is uploaded for cloud transcription; the original video stays in your browser.
Listen back to check quotes, names and numbers. Edit the text, add speaker names yourself and export TXT, DOCX, SRT or VTT before leaving the page.
Bring a phone recording, an exported remote interview or a camera file with spoken audio. For a noisy room recording, reduce background noise before transcription and compare the voice before choosing the cleaned version. If you already have an MP3, you can also use MP3 to text directly; no format conversion is needed just to start an interview transcript.
Free supports original audio up to 25 MB and 10 minutes; Pro supports original audio up to 100 MB and 60 minutes. Video files can be up to 500 MB with the same duration limits. Only prepared audio parts, each up to 25 MB, are uploaded. Browser decoding support and available device memory still matter; large-file compatibility on an iPhone has not been verified.
Before interview transcription, listen to the beginning and a section near the end of your file. Check that both the questions and answers are audible and that the file contains the complete conversation. Keep the original recording unchanged. If you trim a section in another editor, note its starting time: exported timestamps refer to the uploaded file, not the untrimmed interview.
For a research interview, review the question and answer turns and replace identifying details if your project requires it. For an article, compare every quote you plan to publish against the recording. You can edit the transcript segments and include timestamps, then export DOCX for document review or SRT and VTT for subtitles. All four export formats are available on Free and Pro. Text edits are not saved automatically after a reload, so export your reviewed copy before leaving.
Treat interview transcription as a review workflow rather than a finished quote. Make one pass for names, dates, figures and specialist terms, then another for the passages you intend to use. Replay unclear wording instead of guessing; mark an uncertain word as unclear in your working copy. Do not silently turn a hesitant answer into a stronger claim. Export the reviewed document before refreshing or closing the page.
This invented example illustrates a reviewed layout, not actual transcription output. The speaker names and roles have been typed manually.
[00:12] Interviewer: What changed after the first trial? [00:16] Participant: We moved the review to the start of each week.
Use names or participant codes appropriate for your project. Verify each turn against the recording rather than assuming that a new segment means a new speaker.
Timestamps help you find a passage and check a quote against the recording. They do not identify who is speaking. Automatic speaker separation and speaker labels are not supported: listen back and add names or roles in the editable text yourself. Overlapping speech may need extra correction, and this tool does not guarantee a research-ready verbatim transcript.
For research interview transcription, decide on a consistent manual labeling scheme before editing: for example, Interviewer and Participant 01. Keep the same label when a person speaks again, and check short interruptions by listening. A timed segment is not necessarily a complete answer or a change of speaker. Follow your project’s rules for anonymization, filler words and punctuation, and keep any identifying key separately from a transcript you share.
More details about working with your recordings.
Free: up to 10 minutes and 25 MB for original audio. Pro: up to 60 minutes and 100 MB for original audio. Videos can be up to 500 MB, with the same plan duration limits. Each prepared audio part is limited to 25 MB. Your browser must be able to decode the audio; the size limit does not guarantee that every phone can process a large recording.
Yes. Upload an MP4 or MOV with an audio track your browser can decode. Use only recordings you own or have permission to process, and check that any participant consent or project rules allow cloud transcription. The original video stays in your browser; prepared audio is uploaded to private Cloudflare storage and sent to OpenAI. Stored tasks expire after 24 hours and can be deleted sooner.
Yes. The transcript has timed segments, and you can export SRT or VTT subtitle files. Review the words and timing against the recording before using a quote or publishing captions. Fine timing adjustments can be made in your video editor.
No. Automatic speaker detection and speaker labels are not included. You can type names or roles into the editable text after listening to confirm who is speaking. These manual text edits do not add automatic speaker separation or change subtitle timing.
Free: 1 file without an account, up to 10 minutes. Sign in for 30 free minutes each month. Noise removal and WAV download are always free. Your current-month anonymous trial counts toward the 30 minutes. Free minutes reset on the first of each month at 00:00 UTC and do not roll over. Daily transcription capacity: Free shares 1,500 minutes; Pro has a separate 1,500-minute pool, independent of the free pool. Both reset at 00:00 UTC. If your pool is full, try again after the reset; no task is queued for the next day.
Your next idea is already in there.
Let’s get it into words.