Start for free
Pick a plan later or never

Custom wake words, speech-to-text, and voice commands, seamless deployment

Simple, transparent, no BS*


No credit card required
+ Custom Wake Words
+ Custom Voice Commands
+ Voice Activity Detection (VAD)
+ Transcription
+ Real-time Transcription
+ Voice Search
➔ Unlimited voice interactions for 3 monthly active users
➔ 100 hours of transcription and search per month
➔ Custom model training via Picovoice Console
➔ Community support via GitHub

Voice Assistant Starter

Billed annually
➔ Everything in the Free
➔ 1000 monthly active users
➔ Email support

Transcription & Search Starter

Billed annually
➔ Everything in the Free
➔ 10,000 hours per month
➔ Email support

& Scale

➔ Everything in the Free
➔ Unlimited users & hours
➔ Dedicated support

Frequently Asked Questions

Choosing the best voice recognition engine for your voice products

Check out Leopard Speech-to-Text. Speech-to-text (STT) is also known as voice-to-text, automatic speech recognition (ASR), automated speech recognition, audio transcription, or just transcription engine.

Check out Octopus Speech-to-Index. Speech-to-Index refers to audio search, audio indexing, speech indexing, acoustic indexing and acoustic searching.

Check out Porcupine Wake Word. Wake words are also known as trigger words, hotwords and wake-up words. A wake word is a specific form of keyword spotting (KWS). Wake word detection is also known as hotword detection, keyword recognition and keyword detection.

Check out Rhino Speech-to-Intent. It fuses speech-to-text (STT) and natural language understanding (NLU) to infer intents directly from spoken utterances. Similar technologies are known as spoken language understanding (SLU), spoken intent detection, and spoken language intent detection.

Check out Cobra Voice Activity Detection. Voice Activity Detection (VAD) is also known as speech activity detection, speech detection, or voice detection.

Free or Paid: Choosing the best tier for your voice recognition needs

Yes! Anyone can use Picovoice to train voice models and deploy them, even commercially, for free.

Reach out to sales and tell us about your project, use case, volume, and requirements. Someone from the Picovoice team will get in touch with you shortly.

No, you can directly start with the Enterprise Tier by contacting sales with your business requirements.

Transcription & Search Starter Tier grants access to up to 10,000 hours of monthly transcription and voice search with Leopard Speech-to-Text, Octopus Speech-to-Index and Cheetah Speech-to-Text. This package offers all you need to start transcribing audio recordings such as meetings, call center recordings, type-by-voice, closed captioning, and search keywords within recordings.

Voice Assistant Starter Tier grants access to up to 1,000 monthly active users with Porcupine Wake Word, Rhino Speech-to-Intent and Cobra Voice Activity Detection. This package offers all you need to start building voice interfaces (VUIs) such as voice assistants, voice-activated menus, voice-enabled applications or voice-controlled devices.

No, you can use Picovoice engines standalone or in combination. It depends on your needs and use case. Picovoice offers all you need for audio transcription and voice search under the Transcription and Search package with Leopard, Octopus and Cheetah, and for voice interface and assistants under the Voice Assistant package with Porcupine, Rhino and Cobra. Check out the Strategy Guide and Voice Search, Voice Control, Search by Voice and Speech Analytics use cases to learn more.

No, the smallest paid package is the Starter Tier.

Transcribe & Search and Voice Assistant Package Details

Voice Assistant packages are priced based on the number of monthly active users. Monthly active users are tracked via AccessKey. Check out the FAQ to learn more about Picovoice technology and pricing.

Transcribe & Search packages are priced based on the amount of voice data processed per month. Monthly second-based consumption is tracked via AccessKey. Check out the FAQ to learn more about Picovoice technology and pricing.

Allowances and real-time consumption can be tracked on the Picovoice Console profile page.

No. Your allowances reset every 30 days.

By that time, you’ll have contacts from the Picovoice team who will help you out.

About Picovoice and support

Picovoice does not track, collect or store user data to train voice AI models or process voice data. Picovoice, on-device voice recognition technology, processes voice data locally on-device with no access to user data.

While data is processed offline, locally on-device, Picovoice engines call home servers to stay active and report the consumption for billing purposes only.

Picovoice Console, docs and GitHub are designed to help you throughout your development journey and answer your questions. However, if your questions are not answered there, please feel free to create an issue under the relevant GitHub repository with as much detail as possible. If you’re a paid-tier customer, you can reach out to your contact at Picovoice.

Picovoice docs, FAQ, glossary, blogs, Medium posts, and GitHub are great resources to learn about voice AI, Picovoice engines and how to start building voice products.

No BS represents the simplicity and transparency that Picovoice commits to developers. Instead of BS (bad sales tactics), Picovoice
  • builds open-source benchmarks to enable data-driven decision-making,
  • makes its technology available via Free Tier with no product, platform or SDK limitations,
  • streamlines the developer experience via self-service Picovoice Console and modern SDKs.
Picovoice is the end-to-end platform for adding voice to anything on your terms.