Stream audio for live transcription and real-time translation, and watch source-language and translated captions appear together on screen. Sub-second end-to-end. Audio never leaves the device.
On-device live captioning and translation combines streaming speech-to-text with machine translation in a single local pipeline. Most implementations split these across a cloud STT API and a separate translation API, paying network round-trip latency twice and sending audio off the device at every step. Picovoice runs both stages on-device: the streaming STT engine produces source-language captions as the speaker talks, and the translation engine converts them into the target language in the same pipeline, with no audio or text ever transmitted to a third-party service.
Cheetah Streaming Speech-to-Text beats Google Cloud STT in word error rate and word emission latency across all tested languages, and outperforms Azure STT in several benchmarks, per the open-source real-time transcription benchmark — even before it's customized for the use case. It emits words at 590 ms median latency, typically one word behind the speaker. Cheetah requires less compute than any other local engine tested. Result: no tradeoff on accuracy, latency, or privacy, and no minimum hardware requirements.
Zebra, Picovoice's on-device translation SDK, returns up to 120 words per second — far faster than the average 2 words per second a person speaks or the 5 words per second a person reads. The speed does not come at the cost of accuracy: it matches the accuracy of Opus, one of the best-known open translation models. On-device processing also gives Zebra a competitive edge: zero network latency, which can never be achieved by cloud translation APIs such as Google Translate and DeepL.
Broadcasts — whether for a company-wide meeting, a city-wide council discussion, or educational content — require live captioning and translation for accessibility and reach. They need captions that track audio within seconds and survive flaky uplinks. On-device live captioning and translation meets these requirements and captions the same feed in multiple languages simultaneously.
Apple Live Captions and Google Live Captions work well within their own infrastructure. If you are shipping CART captioning or accessibility features in a cross-platform product — including lecture capture systems, hospital communication tools, or public-sector apps — Picovoice gives you the same on-device live captioning capability across Android, iOS, web, embedded, Linux, macOS, and Windows.
A complete working recipe in Python. Open-source on GitHub. Runs 100% on-device.