vsref
Picovoice (Leopard / Cheetah) logo

Picovoice (Leopard / Cheetah) Review

Private, on-device STT SDK for apps and edge devices

Among the 43 speech-to-text tools we track, Picovoice (Leopard / Cheetah) has the 38th-widest language coverage.

See pricing

Facts verified Jul 20, 2026Try Picovoice (Leopard / Cheetah) →

If you sign up through links on this page, vsref may earn a commission; programs exist on both sides of most comparisons, and commissions never change verdicts. How we make money

What we know about Picovoice (Leopard / Cheetah)

This is our verified profile of Picovoice (Leopard / Cheetah), a speech-to-text apis platform - private, on-device STT SDK for apps and edge devices. Every fact about Picovoice (Leopard / Cheetah) below carries the source it came from and the day we checked it.

Picovoice (Leopard / Cheetah) does not publish a public per-minute rate we have been able to verify, so the price figure here stays blank until we can confirm one. We would rather show nothing than a guessed number; if you have a current Picovoice (Leopard / Cheetah) quote, it is the fastest way for us to close that gap.

On capabilities, Picovoice (Leopard / Cheetah) covers word-level timestamps, custom vocabulary / keyterm boosting, and self-host / on-prem option, and does not offer websocket streaming api. Each of those is verified against Picovoice (Leopard / Cheetah)'s own docs or dashboard, not marketing copy.

Placed against the 43 speech-to-text tools we track, Picovoice (Leopard / Cheetah)'s strongest showing is the 38th-widest language coverage - a spread worth weighing against your own priorities.

In total we track 13 verified facts for Picovoice (Leopard / Cheetah) today, each linking the primary source it came from so you can check our work - and vendor claims we have not measured ourselves are labeled as such on the Picovoice (Leopard / Cheetah) fact sheet below.

Reviewed by vsref Editorialfacts verified Jul 20, 2026Methodology →

Fact sheet

Capabilities
Capabilities facts
WER (vendor-claimed)~9.7 % WERJul 20
Languages supported8Jul 20
Speaker diarization✓ IncludedJul 20
Word-level timestamps✓ YesJul 20
Custom vocabulary / keyterm boosting✓ YesJul 20
Compliance & trust
Compliance & trust facts
Self-host / on-prem option✓ YesJul 20
Build experience
Build experience facts
Official SDKsPython, C, Java, .NET, Node.js, iOS, Android, WebJul 20
Supported audio formats3gp, FLAC, MP3, MP4/m4a, Ogg, WAV, WebMJul 20
Websocket streaming API✗ NoJul 20
Commercial
Commercial facts
Model size (parameters)Leopard ~37 MB, Cheetah ~34 MB model filesJul 20
Hardware to self-hostCPU-capable incl. Raspberry Pi 3+; optional GPUJul 20
Project maintenance statusActive - Cheetah v4.1.0 released 2026-06-24Jul 20
GitHub stars669Jul 20

Considering a switch? Best Picovoice (Leopard / Cheetah) alternatives →

Voice agents in this stack

The engine is one layer: a voice agent hears through its transcription engine, but orchestration, telephony, and turn-taking come from the agent platform. Compare the platforms builders pair Picovoice (Leopard / Cheetah) with: Pipecat, OpenAI Realtime API, Twilio ConversationRelay, or the full voice-agent comparison →

Want it done for you?

Picovoice (Leopard / Cheetah) gets you a transcript; the bot that joins the call, labels speakers, and writes the summary is still your build. If that is more pipeline than you want to own, AI meeting notetakers do the whole job end to end: Otter.ai, Fireflies.ai, Fathom, or the full notetaker comparison.