iOS 17 · watchOS 10 · BYOK · No subscription
For engineers in China meetings, expats and conference interpreters who need it to actually work.
Five Pipelines
Each pipeline is a curated ASR→MT→TTS stack — tuned for its situation. No guessing, no tweaking.
The pipeline picker shows latency, quality and cost per stack — so you always know what you're running. Switch mid-session without restarting.
Full pipeline deep-dive →
BYOK — Bring Your Own Key
"You bring the keys. SpeechKey is the cockpit.
Every API call goes directly to the provider you chose — we have no servers, no middleman markup, no surveillance pipeline."
The Providers screen shows all configured providers, active pipeline status and connection state. Switch provider in real time — no restart, no interruption.
Set up API keys →
Privacy
SpeechKey has no backend. We don't store, see, or sell your conversations. No account, no email, no tracking.
Read the full privacy policy →SpeechKey runs entirely on your iPhone and Apple Watch.
No email, no signup, no password.
Never transmitted anywhere except to your chosen provider.
Exportable, deletable, never uploaded.
We have no data to delete because we have no data.
Declared with CA92.1 reason. No tracking APIs.
Pricing
"A Pro tool at indie prices,
not a SaaS dressed up as an app."
¥99 in China · AI usage costs go directly to your provider