Picovoice

What it does

Picovoice offers on-device AI SDKs for voice, language, and vision. The product line includes wake word, speech-to-text, text-to-speech, speaker recognition, translation, and OCR. Each product ships as a self-contained SDK that runs locally, so there are no cloud API calls and no data leaving the device.

Picovoice is a set of on-device AI SDKs for voice, language, and vision. The homepage lists fifteen production-ready products, including wake word, speech-to-text, text-to-speech, speaker recognition, translation, and optical character recognition. Each product ships as a self-contained SDK, and the site states there are no cloud API calls, no network latency, and no data leaving the device. The company also provides app blueprints with demo and open-source code for use cases such as live captioning, voice assistants, and document QA. A free trial is available without a card, and enterprise buyers can talk to sales. The homepage does not publish pricing, so cost details require a sales conversation. The site makes a performance claim about voice activity detection compared with Silero, but no independent benchmarks are provided. Device support and accuracy should be checked for your own audio conditions.