01Why do I need my own API keys?
Because it's the only honest model that lets us run no servers. If SpeechKey managed the keys for you, we'd be another company sitting between you and your data — with a markup on every API call, a server logging everything, and vendor lock-in. BYOK means you pay OpenAI and iFLYTEK directly at their actual rates; no middleman, no hidden margin, no surveillance pipeline. SpeechKey is the cockpit layer. The engines are yours.
02What does it actually cost per hour?
It depends on the pipeline:
- China-Only
- ~$0.60 / h
- Standard
- ~$1.20 / h
- Pro Live
- ~$6.00 / h
- Realtime
- ~$9.00 / h
At Standard, five hours of meetings is roughly $6. These are estimates for continuous speech activity. Your provider dashboard shows exact usage.
03Does it work in China without VPN?
Yes — with the China-Only pipeline. It uses iFLYTEK for speech recognition, DeepSeek for translation and MiniMax for voice output, all on servers within mainland China. No traffic goes to OpenAI or other internationally blocked endpoints.
All other pipelines (Standard, Pro Live, Realtime, Watch-Modus) require VPN because they use international providers such as OpenAI.
04Is my audio sent to your servers?
No. SpeechKey has no servers. Your audio goes directly from your iPhone to the API endpoint of the provider you chose — iFLYTEK, OpenAI, etc. SpeechKey is the client layer on your device. We don't see your conversations, can't log them, and have no means to store them because we operate no infrastructure for that.
What the provider does with the data it receives is governed by their privacy policy. Review the privacy policy of your chosen provider.
05Which iPhone do I need?
SpeechKey requires iOS 17 or later. The Apple Watch companion requires watchOS 10 or later. There's no other hardware requirement — any iPhone capable of running iOS 17 works.
06Will it work for my dialect / industry terminology?
Mandarin (Putonghua) — solid across all pipelines. Cantonese — recognised by iFLYTEK ASR, but with lower accuracy particularly for fast informal speech. Other Chinese dialects — best tested against your specific audio environment.
For technical / industry terminology: the Terminology Manager injects your own glossaries into the MT system prompt at runtime — proper nouns, project codes, brand names and automotive codes stay consistent.
07Can I use it offline?
No. All five pipelines require an internet connection for ASR and/or MT. SpeechKey is designed for professional environments where Wi-Fi or mobile internet is available.
08What about languages other than Chinese?
SpeechKey is optimised for ZH↔EN and ZH↔DE. Other language pairs are technically supported — iFLYTEK recognises multiple languages and GPT-4o translates into virtually any language — but the pipeline configurations, latency tuning and Terminology Manager templates are optimised for Chinese. For other language pairs, quality may vary significantly.
09Why no subscription?
Because we operate no servers. A subscription model typically requires backend infrastructure managing user accounts, payments, licensing and (usually) media routing. We deliberately didn't build that. Without servers there are no ongoing operational costs that would justify a subscription — so we don't charge one.
What you pay: 9,99 € once for the app. You also pay your provider directly for API usage — but that's not a SpeechKey fee, that's the actual cost of the AI infrastructure you're using.
10What happens when GPT-5 launches?
The pipeline configurations will be updated to include new models as they become stable and suitable for live-translation latency. Updates are included in the app price. You don't need to do anything except install the app update.
BYOK also means: when OpenAI releases a better model at the same cost, you benefit automatically — no price increase, no tier upgrade required.
More questions? Email support@vaitl.ai.