TwoTranscribe
Interview transcription with AI
Turn a recorded interview into a transcript you can read, check and work with. Enable speaker recognition to distinguish contributions, then export the text to Word for your research or article.
A working transcript for research and journalism
Use TwoTranscribe after a participant interview, customer conversation or recorded Q&A. Read the discussion as text, look for a phrase and return to the recording to check a quotation. A transcript gives you material to analyse; you still choose the themes and decide which passages to publish.
Keep the voices apart
Choose speaker recognition before uploading. Speaker labels help separate questions from answers, but they do not establish a person’s identity. Check the labels against the audio and edit the speaker names in the result when needed. Overlapping speech and a distant microphone can make attribution less reliable.
Prepare quotes and a Word document
Download DOCX to annotate passages in Word, PDF for reading or TXT for your own analysis workflow. With speaker recognition, start/end times for speech segments help you locate a quote in the recording. Check names, figures and quotations against the audio before citing them. Review speaker attribution again when a new part of a longer interview begins.
From upload to download
Upload your file
Open the workspace, choose your recording and select its spoken language. Set an optional translation language separately and enable speaker recognition if useful.
Review the cost
Check trial availability and the price for the remaining recording. You can pay once without an account. To use tokens, log in with enough balance.
Start processing
Confirm full processing through the payment or token route shown. Follow progress in the workspace; processing time depends on your recording and the queue.
Review and download
Check important passages, then download TXT, PDF or DOCX. A completed translation is a separate result from the source transcript.
An interview about a customer’s workflow
[Speaker 1] (12.00s – 16.50s)
What happened when you needed the quote again?
[Speaker 2] (18.00s – 22.50s)
I remembered the subject, but not which recording it was in.
[Speaker 1] (25.00s – 29.50s)
What would help you find it?
[Speaker 2] (28.00s – 32.50s)
A transcript I can search for the phrase.
Files, languages and limits
- Upload formats
- MP3, WAV, M4A, MP4, MPEG, MPGA, OGG, OGA, FLAC and WEBM. Video files are processed for audible speech.
- Maximum upload
- 64 MB per file on this installation, with up to 120 minutes of audio. Split longer recordings and choose a smaller file if it exceeds the upload limit.
- Text downloads
- TXT, PDF and Word (DOCX). A completed translation and the source transcript are separate downloads.
- Speakers and time references
- With speaker recognition, the text includes speaker labels and start/end times for speech segments. Labels may differ between parts of a longer recording; check who is speaking. Without speaker recognition, you receive continuous text without these segment times.
- Quality and keeping results
- Check names, figures, specialist terms and quotes against the recording. Unused uploads expire after 24 hours, source audio 24 hours after delivery and guest results after seven days. Download results you want to keep.
- Availability
- Each account or guest can upload up to five new recordings per UTC day. Shared daily and capacity limits can also restrict uploads temporarily. The workspace shows a message when a limit is reached.
Available source and translation languages
Spoken language: automatic detection or Arabic, Bulgarian, Czech, Danish, German, Greek, English, Spanish, Estonian, Finnish, French, Hungarian, Indonesian, Italian, Japanese, Korean, Lithuanian, Latvian, Dutch, Polish, Portuguese, Romanian, Russian, Slovak, Slovenian, Swedish, Turkish, Ukrainian, Chinese, Norwegian.
Optional translation languages: Arabic, Bulgarian, Czech, Danish, German, Greek, English, English (British), English (American), Spanish, Estonian, Finnish, French, Hungarian, Indonesian, Italian, Japanese, Korean, Lithuanian, Latvian, Norwegian, Dutch, Polish, Portuguese, Portuguese (Brazilian), Portuguese (European), Romanian, Russian, Slovak, Slovenian, Swedish, Turkish, Ukrainian, Chinese, Chinese (Simplified), Chinese (Traditional).
What does it cost?
You can upload, use an available trial, pay per recording and download the result without an account. There is no monthly subscription.
Try up to 5 minutes, once per visitor per day measured in UTC, while the shared trial budget is available. A recording within the available trial is free. If the trial is used up or unavailable, you can pay for the full recording.
For a one-off payment, pay €0.15 per remaining minute, with a minimum payment of €1.95 whenever there is a paid remainder. We calculate per second and round the final amount to cents. Speaker recognition and one optional translation are included in this rate.
A 60-minute recording costs €8.25 with an available 5-minute trial, or €9.00 without a trial. Your exact recording price is shown before payment; check the final total at checkout.
Prefer to use tokens?
Log in to use an existing token balance. Each remaining second of audio uses one token; speaker recognition adds 33%, rounded up. One optional translation is included. The workspace shows the exact balance needed after upload. A token purchase creates an account. This payment route is separate from the one-off recording price above.
- Enterprise: € 85.00 · 50,000 tokens
- Starter: € 3.50 · 1,000 tokens
- Standard: € 12.50 · 5,000 tokens
- Pro: € 22.00 · 10,000 tokens
Frequently asked questions
Can I use this for qualitative research?
You can use the transcript as material for coding and analysis. Check it against the recording and follow your research method for pauses, omissions and quotations. Automatic text is not a certified verbatim record.
Will TwoTranscribe name the interviewer and participant?
Speaker recognition distinguishes voices with labels. Check those labels and rename speakers yourself where appropriate; it does not identify people by name.
Can I download the interview in Word?
Yes. Download the original transcript as DOCX and open it in Word. TXT and PDF exports are also available. A requested translation is a separate result.
How should I record an interview with two speakers?
Keep both voices close to the microphone, reduce background noise and let each person finish speaking. Listen to a short section before recording the full interview.
Start with your own recording
Upload your file and review the available next step and cost in the workspace.