“Integrating a wake word engine is expensive, involves upfront cost, and takes time. Although affordable for big companies this can be a show-stopper for startups who want to join the voice-enabled revolution.” This is a snippet of Picovoice’s first article when we launched Porcupine (co-incidentally our first product). Fast forward a few years and we are even more committed to the democratization of voice AI. Before wrapping up 2021, we take a significant leap towards this goal.


Anyone with an email address, GitHub, LinkedIn or Google account can train on-device speech models and deploy them, even commercially, for free. This includes Leopard Speech-to-Text (file-based), Cheetah Speech-to-Text (streaming), Octopus Speech-to-Index (voice search), Porcupine Wake Word, Rhino Speech-to-Intent, and Cobra VAD engines. See below for details.

Free Tier for Everyone

No credit card is required. You don’t need to ever convert to a paying customer if you are working on a small project, personal or commercial. You can use all capabilities of Picovoice Console to train voice models and deploy them using any of our SDKs to your target device.

Commercial Explorations

Voice Assistant Starter, unlimited voice interactions with Porcupine, Rhino and Cobra for up to 1000 monthly active users (devices), is for $899 a month all inclusive — no hidden fees. Transcription and Search Starter, 10,000 hours of transcription (real-time or file based) and indexing for unlimited voice search, is for $999 a month all inclusive — no hidden fees. After years of helping enterprises bring their voice products to market, we believe this is the right plan for the beginnings of a commercial effort.

Closing Remarks

If you like to learn more, check the pricing page for details of the above plans. Ready to build? You are just a few clicks away! Docs is a great start.