Where your data is processed
All your content data, meaning audio, transcripts, account details and payments, is processed and stored in the Netherlands. Murmel is a product of The AI Factory B.V. in Amsterdam and works with these providers:
| Provider | What for | Location |
|---|---|---|
| Scaleway | Servers, storage of audio and transcripts, email | Netherlands |
| Leafcloud | GPUs that run the speech model | Amsterdam |
| Mollie | Payments | Netherlands |
| Only measuring website visits and ads | US / EU |
The speech model is Murmel's own. So your audio is not sent to an external AI service to be converted. Leafcloud's GPUs sit in buildings in and around Amsterdam, where the servers' waste heat is reused.
Who can access it
- You. Your recordings and transcripts are in your own account, behind your own login.
- Murmel, only where needed. Access to data is limited to authorised staff, for example to fix a problem.
All connections use HTTPS/TLS. Passwords and API keys are only stored hashed. In case of a data breach that poses a risk to your rights, Murmel reports it to the Dutch Data Protection Authority within 72 hours and informs you as soon as possible.
How long your data is kept
You choose, in the Studio settings or on your profile page:
- Keep until you delete it yourself. The default for new accounts.
- Delete automatically after 1, 7, 30 or 90 days or after 1 year. Both audio and transcript are then removed.
- Delete yourself, per recording or several at once in the Recordings tab. You can also delete your whole account with all its data.
For accounts created before 21 October 2026 the earlier setting still applies: without "Keep longer", audio and transcripts are deleted after 7 days. Payment records are kept for 7 years because of the legal retention requirement.
No training without consent
Your recordings are not used to train the model unless you opt in yourself with the Training data setting. If you do, the recordings you offer are kept until you withdraw consent or delete the recording.
Through the API: store nothing
If you connect Murmel to your own software, store=false means nothing is kept. Audio and transcript are deleted as soon as you have fetched the result, and at the latest one hour after the transcription finished. See the API.
For organisations: GDPR, processing agreement and DPIA
If you record clients, citizens, patients or colleagues, you (or your organisation) are responsible for a valid legal basis, for example consent. Murmel processes the audio on your behalf. For organisations, Murmel provides:
- A data processing agreement (DPA) tailored to you.
- Help with your DPIA: answers to your privacy officer's questions about data flows, retention periods and sub-processors.
- Murmel Teams: one admin, shared minutes and one invoice, with payment by invoice and an SLA by arrangement. See Murmel for teams.
Request it through the contact page. For municipalities and other public bodies there is more on transcription for government.
The full legal text is in the privacy policy. It also lists your rights under the GDPR, such as access and erasure.
Frequently asked questions
Does my audio go to the United States?
No. Audio, transcripts, account details and payments stay in the Netherlands. Only measuring website visits and ads runs through Google, and that never touches your recordings.
Is my recording used to improve the model?
Only if you opt in yourself with the Training data setting. Not by default.
May I record a conversation and have it transcribed?
Often yes, but you are responsible for a valid legal basis. Ask at the start whether everyone agrees to the recording, especially when personal data is discussed. This is not legal advice; if in doubt, ask your privacy officer.
Can I get a data processing agreement?
Yes, for organisations and teams. Get in touch with the name of your organisation and what you want to transcribe.