Transcribing interviews for journalism: getting quotes right

A misattributed or out-of-context quote can damage an entire piece. Anyone using interviews journalistically needs more than just any transcript — accuracy, traceable timestamps, and careful handling of the recording are part of editorial due diligence.

How do you transcribe an interview?

Record each speaker clearly, create an AI draft with speaker diarization, rename the speaker labels, then verify names, numbers, and every quotation against the audio. Timestamps make the final check faster and preserve a traceable path from published text back to the recording.

Before recording: consent and transparency

Make it clear at the start of the conversation that you're recording and what the material will be used for. Recording someone without consent can be unlawful depending on jurisdiction — for journalistic interviews, the interviewee's consent isn't a formality, it's a precondition.

From raw transcript to a print-ready quote

• Automatic transcription gives you the full wording including slips and filler words — that's the starting point, not the finished quote
• Per-section timestamps let you re-listen to a passage later and verify the exact wording
• When editing for readability (removing fillers, smoothing broken sentences), the meaning must not shift
• Check names, titles, and technical terms against a second source before publishing — AI transcription can miss proper nouns

Keeping multiple speakers straight

In interviews with multiple participants or interjected questions, speaker diarization automatically assigns sections to different voices. Before publishing, it's worth checking the first few minutes to match the automatically assigned speaker labels to real names — automatic attribution can slip, especially with short exchanges or overlapping speech.

Retention and traceability

In case of a follow-up question or dispute, it's sensible to keep the original recording and transcript for as long as your editorial due-diligence practice requires. Sprechverlauf doesn't permanently store transcripts server-side — for your own documentation, export the result directly as Word or PDF.

What Sprechverlauf offers for interviews

Per-section timestamps for quickly finding quotes again, speaker diarization for conversations with multiple people, and an export format you can drop straight into editorial workflow — free, no account, no per-minute pricing.