Users and GDPR
Voice recordings, transcripts, and chat logs are personal data. Keeping them in-region keeps your GDPR story simple and your DPAs boring, in the good way.
// EU-compliant AI
AI features that keep data in the EU.
Voice and chat features are personal-data pipelines. We design and build AI stacks where every stage, from microphone to model and back, runs in EU regions, and we document it so you can prove it.
// Why it matters
Residency is the question you'll be asked.
Enterprise buyers ask
"Where does the data go?" is now a first-call procurement question. "It never leaves the EU" is the answer that shortens security reviews.
The regulatory direction
Between GDPR enforcement and the EU AI Act phasing in, region-pinned architecture is the safe long-term default. Cheaper to build now than to retrofit later.
// How we build it
Map, pin, prove.
The same approach whether we audit an existing AI feature or build a new one from scratch.
-
Map
Every AI data flow gets drawn: what leaves the device or browser, where it lands, what is stored, and for how long.
-
Pin
Each stage runs on providers with EU regions: Azure Speech for STT and TTS, Vertex AI or Bedrock EU regions for the LLM, EU-only storage and logging. Self-hosted open-weight models when no managed option fits.
-
Prove
You get the data-flow diagram and the configuration as deliverables: the artifact your DPO or counsel actually reviews, instead of a verbal promise.
// Shipped, not slideware
An AI companion device
EU residency was a hard requirement, not a nice-to-have: speech on Azure STT/TTS, the LLM on Google Vertex AI, both pinned to EU regions, streaming in real time to a physical device.
- ESP32
- Flutter
- Laravel
- WebSocket
- Azure STT/TTS
- Google Vertex AI
- EU residency
// Get in touch
Have something to build?
Tell us what you're working on. We'll reply within one business day.
NDA on request · Full IP transfer to you.