Where your voice goes, exactly
What is confirmed in the prototype, the privacy choices already made and the rules still to be validated before launch.
The journey we are preparing
- 01Activation on your device
- 02Encryption in transit
- 03Transcription on European infrastructure yet to be selected
- 04Goal: deletion when processing ends
- 05Insert text into the active application
The maximum technical duration has not been validated yet. It will be measured, governed and published before the service opens.
Four choices to confirm
- Infrastructure
- European
- Provider and regions to be confirmed
- Treatment
- European Union targeted
- Exact regions not published
- Audio retention
- Policy being validated
- Maximum duration published before launch
- Training
- Off by default
- Deliberate activation in settings
Details of treatments
Hosting
We want to operate Speakey on European infrastructure. The provider, regions, backups and associated data flows will be selected and published before the first pilots.
- Planned criteria: European location, encryption and auditability.
- The processing regions will be detailed after validation.
Voice processing
The audio stream is encrypted in transit to produce text. The product goal is to delete it when processing ends in most cases. The maximum technical duration and rare retention needs will be defined before launch.
Speakey does not use your dictations to train its models by default. This requires deliberate activation in app settings and can later be disabled.
RGPD
Depending on the context, Speakey acts as controller for its service or as processor for the client organisation. Roles and instructions are set out in the final contractual documents.
Your rights
Access, correction, deletion and portability procedures will be documented before the product opens.
Contact personal data
The dedicated channel for personal data requests has not been published yet.
Security
- Encryption of data in transit is planned in the target architecture.
- No certification is claimed until obtained.
Processors
No provider is presented as a processor for the future service before being selected. The contractual list, roles and locations will be published before any voice processing begins.
| Processor | Role | Location |
|---|---|---|
| To be selected | Hosting and compute resources | France |
Documents
The data processing agreement, privacy policy and contractual list of processors will be made available once validated. No draft document is presented as final.
- Data processing agreement (DPA)Publication forthcoming
- Privacy policyPublication forthcoming
- List of processorsPublication forthcoming
Would you like to contribute to our privacy choices or prepare a future pilot?