
by Picovoice
Picovoice provides on-device AI SDKs for voice, language and vision, including wake word, speech-to-text, text-to-speech and speaker recognition. Everything runs locally with no cloud API calls.
APIFreemium
Step 1: Start a free trial and get an access key.
Step 2: Try the live demos and code snippets for the product you need.
Step 3: Use Picovoice Console to train a custom wake word, voice commands or speech model.
Step 4: Export the model and add the SDK to your app in your preferred language.
Step 5: Combine several SDKs, such as wake word and speech-to-text, for an end-to-end feature.
Step 6: Talk to sales for enterprise support when ready for production.
No. The SDKs run on the device, and the docs describe voice recognition as entirely offline.
The site lists 15 products across voice, language and vision, such as wake word, speech-to-text, text-to-speech, speaker diarization and OCR. Each ships as a self-contained SDK.
Yes. Picovoice Console supports training branded wake words, voice commands and custom speech models in the browser.
The site says its SDKs run on mobile, web, desktop and embedded devices.
Hume AIHume's Audio Decisions API lets you ask plain-language questions about an audio file and get back probabilities estimating how human raters would answer.
Murf APIMurf API applies AI to AI voice generation API for producing natural text-to-speech audio in applications and automated content workflows.
Speechmatics APISpeechmatics API applies AI to speech recognition API for realtime and batch transcription across accents, languages, and enterprise audio.