What is live speech translation?
Live speech translation listens to speech in one language and produces text or speech in another as the conversation happens — near-realtime, not after the fact. Modern implementations pair speech recognition with translation models; the meaningful differences are latency, accuracy, and where your audio is processed.
How it works
A recognition model converts speech to text in the source language; a translation model renders it in the target language; optionally a voice synthesizes the result. Doing this continuously, with partial results updating as the speaker continues, is what separates live translation from transcribe-then-translate.
The bring-your-own-key model
Most translation apps route your audio through the vendor's own cloud, bundled into a subscription. An alternative model: the app connects directly to an AI provider using your own API key — such as your Gemini or OpenAI key. You pay the provider only for what you use, no separate account or subscription exists, and the app vendor never sees your audio at all. The trade-off is a one-time setup step: obtaining a key.