The researcher's dilemma
You are running a study involving sensitive data. In-depth interviews with patients, refugees, abuse survivors, or whistleblowers. The recordings contain names, medical details, locations and deeply personal stories. Your ethics board approved the study under strict conditions: the data must be handled confidentially and processed anonymously.
Now you need to transcribe those interviews.
Most transcription services ask you to upload audio to their servers. But where are those servers located? Who has access? Is the audio used to train AI models? And how does that align with the promises you made to your participants?
Why standard transcription services fall short
Most transcription platforms are built for convenience, not confidentiality. They process audio on servers in the United States, retain files for days or weeks, and sometimes employ human reviewers for quality assurance. That means real people may listen to your sensitive interviews.
For researchers operating under GDPR requirements and ethical protocols, this is unacceptable. The General Data Protection Regulation requires that personal data is processed only with a valid legal basis and with appropriate technical and organisational safeguards. Transfers outside the EEA face additional scrutiny, particularly since the Schrems II ruling tightened the rules on transatlantic data flows.
What researchers actually need
A transcription solution for academic research must meet four requirements.
1. Data location and jurisdiction
Processing within the EU, on servers governed by European privacy legislation. No transfers to the US or other third countries.
2. Encryption
Data must be encrypted both in transit (TLS 1.3) and at rest (AES-256). Ideally, even the service provider cannot read the content.
3. Data minimisation
Audio is automatically deleted immediately after processing. No long-term storage, no use for model training, no access by the provider's employees.
4. Local processing option
For the most sensitive data, there must be an option to run the transcription entirely on the researcher's own device, with zero data leaving the machine.
How Scribeer meets these requirements
EU-hosted cloud with encryption
Cloud mode processes audio on servers in Frankfurt, Germany. The connection is encrypted with TLS 1.3 and transcripts are stored with AES-256 encryption. Audio is automatically deleted immediately after processing. Data is not used for model training and is not accessible to staff. A data processing agreement is available on request.
Local processing for maximum confidentiality
Private Pro mode runs the transcription model entirely on your own computer using WebGPU. Audio never leaves your device. There is technically no way for Scribeer or anyone else to access the data. After a one-time model download, it works fully offline. This is the level of protection that satisfies the strictest ethical protocols.
GDPR compliance
- Processing within the EU (Frankfurt, Germany)
- AES-256 encryption of stored transcripts
- TLS 1.3 for data in transit
- Automatic deletion of audio immediately after processing
- No use of data for model training
- Data processing agreement available
- Option for fully local processing (zero data transfer)
Practical considerations for your research protocol
If you plan to include transcription software in your research protocol or ethics board application, these are the points to address:
- Describe which processing mode you are using (EU cloud or local) and the rationale.
- Reference the provider's data processing agreement.
- Document the encryption and deletion safeguards.
- Explain how transcripts will be anonymised before analysis.
- Note that the provider has no access to transcript content (for local processing) or that access is technically precluded by encryption (for cloud).
What it costs
Cloud transcription starts at €5.99 for 90 minutes of audio. For larger research projects, team subscriptions are available at a fixed monthly rate. Local processing via Private Pro costs €29.99 per month for unlimited transcription, with a team variant available for research groups.
Many universities reimburse software costs for research. Check whether your faculty or research group has a budget for transcription tools.
Start with a pilot
Transcribe one interview as a pilot before processing your entire dataset. Assess the quality, test the editor, and bring the result to your ethics board as an example of the output. It saves discussion later.
Try it yourself: scribeer.io — 50 free minutes, no credit card required.