Parley Voice FAQ: setup, billing and privacy

Qwen and Doubao currently support live voice cloning, so you do not need to record a voice sample or create a clone in advance.

Your microphone goes to the meeting in their language; their meeting audio plays to you in yours. Venue mode turns one input into one or several languages, each on its own speaker. Both include bilingual captions.

Yes. Zoom, Teams and Meet each ship native speech translation behind platform licences or add-ons, with different language lists. Parley is a desktop app that works across meeting apps through the microphone and speaker settings, with prepaid per-minute billing. For licensing gates and language coverage — including Meet without Chinese as of 2026-10-12 — read the full comparison (Zoom, Teams and Meet native translation vs a desktop app).

No. Your existing audio gear and a Windows or Mac computer are enough. Sending translated speech into a meeting needs virtual audio routing. Managed translation has no recurring subscription.

Choose one of four desktop versions. Review the requirements and setup guide, then check your microphone, headphones and meeting audio before purchasing.

One hour is billed as one 60-minute session at the selected model's per-minute rate. Two directions or several outputs do not double it, and silence while connected counts. See the fixed-cost table for current figures.

Billing counts actual milliseconds once the service connection is ready, silence included. Short sessions aren't rounded up to whole minutes. Two directions cost one session charge. See current pricing.

No. The personal API licence connects your own provider account, which bills usage separately. A licence never adds wallet credit. See the personal API guide.

Not all of it. Selected captured audio is processed by the cloud service; caption history is stored locally. Read the audio and data handling overview.

No. Sign in and add credit to use Parley's managed service. A personal API licence lets you connect your own supported provider in the desktop app. This website doesn't receive model API keys.

No. Choose a compatible built-in voice for the translation service you're using.

Language support depends on the selected model, mode, and voice. Input and voice output coverage may differ. Compare models →

Sending translated voice into a meeting generally needs a virtual audio device. In-person mode uses physical input and output directly.

Yes. Parley translates both ways: what you say is translated for the other side, and what they say is translated for you. To take one direction out of the translation path, change the devices in your meeting app's audio settings. To stop translating your voice, set the meeting app's microphone to your computer's own physical microphone instead of Parley's virtual microphone, so the meeting picks up your original voice. To stop translating their voice, set the meeting app's speaker or output to your own headphones or speakers instead of the virtual audio device, so you hear the original meeting audio. Switch back to the virtual devices at any time to resume translation. While the translation session stays connected, billing still counts the connected time, so turning off one direction does not lower the cost.

Windows installers are currently unsigned, so the system may warn about an unknown publisher. You can continue through the install prompts; we plan to add signing later. Mac requires macOS 14.5 or later.

Ear return lets you hear, in your headphones, how your own words were translated while you speak. It is on by default at about 15% mix, and you can turn it off or adjust it.

Audio is used for live translation; retention and deletion details are in the privacy policy. In personal API mode your keys talk to the provider directly, and the provider bills usage.

Yes. You can follow bilingual captions without sending translated speech. Meeting and venue modes both support captions, and you can export history afterward (subject to the current desktop features).