The interview profile uses punctuation to group questions and source responses while keeping the complete timed transcript available for review.
Question and source-response sections make a long transcript easier to scan.
The complete transcript keeps speaker labels and timestamps beside the spoken text.
Export DOCX, TXT, or JSON after checking names and quotes against the recording.
You will create an account on the next step. The file and its result stay with that account.
When the text is ready
Ready · 4:02 · Interview
interview_2026-08-12_11-38-00“What made you decide to change direction that year?”
“We had already lost two key clients, so waiting was not really an option anymore.”
“How did the team react?”
“Slower than I expected, most people needed to see it work first.”
One flow
Use an audio or video file up to five hours. One recording becomes one interview result.
Use transcript timestamps to find a passage in the recording and edit the text when a name or quote needs correction.
Take DOCX or TXT into the article, or JSON if another tool needs the structured result.
Questions
Transcript segments ending in a question mark become questions; the following segments form source responses. Review the groups rather than treating them as automatic role detection.
Yes. Use the timed transcript and player to find the passage, listen again, and correct the transcript text before publishing.
The interview profile recommends DOCX, TXT, and JSON. Export after the names and quotes in the transcript have been checked.
The minutes profile keeps the recording transcript and analysis annotations together in Studio. Summaries, decisions, and actions are drafts, not a substitute for review.
The lecture profile divides transcript segments into balanced, numbered chapters and keeps coverage through the end of the recording. It does not infer a syllabus.
The subtitles profile turns each timed transcript segment into an SRT or VTT cue. You can correct cue text in the transcript, while the segment timing stays fixed.
The call profile uses question marks and a limited risk-term list to build separate candidate sections. It does not connect each risk to a question or judge the speaker's intent.