Use case
Interview Transcription — from recording to quote-ready transcript
GDPR-compliant transcription for journalism, qualitative research and HR — data stays in Germany.
Interviews are often the source that carries the story — and that's exactly where data privacy matters most. DeepScript hosts all data on its own servers in Germany, no US subprocessors. Word timestamps make every quote traceable; speaker diarization automatically sorts interviewer from interviewee. For on-site two-speaker interviews accuracy stays above 95 % on Premium, also for DACH dialects.
Recommended setup
- Model
- Premium (€0.27/hour)
- Language
- Premium for top accuracy, especially with accents and non-standard speech.
- Custom Vocabulary
- Proper nouns, industry terms, company and place names. For research interviews: theory names and methodology terms.
- Export
- DOCX for quote review, JSON for qualitative analysis tools (MAXQDA, NVivo), SRT if the interview is also published as video.
How it works
- 1
Upload the recording
From any app or recorder: Voice Memos, Zoom cloud recordings, Audacity, digitised tapes. Up to 5 GB per file.
- 2
Set the speaker count
For one-on-one interviews set num_speakers=2. For focus groups give min/max. Improves separation when speakers sound alike.
- 3
Activate custom vocabulary
Add names + technical terms up front for measurably better accuracy. Proper nouns are the most common error source.
- 4
Polish in the editor
Audio stays synced to the transcript — click any word to jump there. Makes verification and context-check fast.
Worked example
Example: 90-minute qualitative research interview, DE dialect (Swiss German), Premium model. Cost: €0.41. Processing time: 4-6 min. Output: word-accurate speaker diarization, JSON for MAXQDA import.
Frequently asked questions
Are interviews stored GDPR-compliantly?
Yes. Own servers in Nuremberg, no US cloud, EU-to-EU DPA online signable. The right setup for university research and journalism with source protection.
Can audio files be deleted automatically?
Yes, that's the default — audio + transcript are deleted 30 days after upload unless Pro is active (then permanent). You can also delete earlier manually.
Does this work for phone interviews?
Yes, as long as you record the audio. Phone audio is bandwidth-limited (300-3400 Hz) and harder to transcribe — Premium model is noticeably better here.
Can I batch-process interviews via API?
Yes. The REST API supports parallel uploads, webhooks deliver completion events, the MCP endpoint exposes the archive to AI agents. OpenAPI spec at [/docs](/docs).
Try it now?
Three transcriptions free, no credit card. Data stays in Germany. Three minutes from sign-up to finished transcript.