Privacy
Local-first by architecture, not just by promise.
The privacy story should be legible to buyers: what stays on the device, what is optional, and why local processing matters for both trust and speed.
Core principles
The trust story needs concrete language.
This page translates architecture choices into buyer-readable product claims without hiding behind generic policy copy.
Local by default
The default experience should not require sending spoken work to a server.
- Core transcription runs on-device by default.
- Cloud is opt-in when a team explicitly wants it.
Clear boundaries
Say exactly what is processed, what stays local, and what is optional.
- Explain what leaves the machine and what does not.
- Remove ambiguity from the trust story.
Metadata-only telemetry
Product analytics should help the team improve the app without collecting dictated content.
- No transcript payloads in observability tooling.
- Keep the privacy promise technical and specific.
Safer daily workflows
Privacy only matters when it stays compatible with the speed and convenience people want.
- Keep latency low enough for real insertion.
- Avoid the compliance anxiety cloud-only tools create.
Comparison
Why local-first matters.
Privacy is only convincing when the site explains the practical difference between a local-first posture and a cloud-first default.
Cloud-only voice tools
Your raw spoken work leaves the device by default, and the trust story depends on vendor promises and policy documents.
Zoryn's local-first posture
Core processing can stay on the machine, cloud can stay optional, and the privacy explanation can stay concrete.
Details
The boundaries should be clear.
If a buyer needs to explain Zoryn internally, this is the part of the site they should be able to copy into that conversation.
Architecture
Privacy is a product capability
The site should explain privacy in terms of actual system behavior, not marketing fog.
- Default mode favors local processing.
- Cloud providers are explicit options, not hidden dependencies.
- API keys belong in secure OS storage, not plaintext configs.
Telemetry
Separate observability from dictated content
Teams need to know whether product analytics can see what they said.
- Usage metadata helps the product team improve the app.
- Transcript text and dictated content should stay out of observability pipelines.
- Trust becomes easier when boundaries are specific.
Reliability
Local processing improves more than privacy
Latency and availability improve when the product does not depend on a round trip for every utterance.
- Local-first lowers the floor on response time.
- The product keeps working in more real-world situations.
- The trust story and the speed story reinforce each other.
Install
Install it, then evaluate the privacy story against the actual product.
The trust model should hold up in the app itself, not just on this page. Start with the local-first default and inspect how the product behaves in your own workflow.