OpenViSTAI · Agents · Blockchain · Crypto
Submit a project
Menu
Audio & Voice AI

Cheetah

Cheetah is Picovoice’s deep-learning library for streaming speech-to-text, processing voice locally on-device to produce transcription. It provides SDKs and demos across programming environments including Python, C, mobile, web and desktop development.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Cheetah — OpenViST project profile in Audio & Voice AI.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Consider Cheetah when an application needs streaming transcription and local voice recognition, and its supported languages and SDK targets fit the project.

Who it is for

  • Developers adding live transcription to an application.
  • Teams building speech input that should process audio on-device.
  • Developers evaluating speech-recognition SDKs across supported environments.

Documented uses

  • Convert streaming audio into text.
  • Add speech input to an application using a documented SDK or demo.
  • Choose a supported language and development environment for a transcription feature.
Before you choose

Requirements & limits

Picovoice requires a valid AccessKey to deploy its SDKs. Internet connectivity is needed to validate the key, although voice recognition runs offline. The repository identifies the project as open source under the Apache-2.0 license.

  • The repository lists six supported languages; additional languages are offered to commercial customers case by case.
  • AccessKey validation requires internet connectivity even though speech recognition runs offline.
  • SDKs and demos are listed for Python, C, iOS, Android, Flutter, React Native, Java, .NET, Node.js and web.
  • Documented platforms include Linux, macOS, Windows, Android, iOS, major browsers and Raspberry Pi models 3, 4 and 5.

Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.

Source-based practical guide

Run the Cheetah Python microphone demo

Install the Python demo package and start its microphone transcription demo.

Before you start

  • Python package installer pip3.
  • A valid Picovoice AccessKey obtained from the Picovoice Console.
  • Internet connectivity to validate the AccessKey.

Steps

  1. Install the demo package with the command below.

    Documented command
    pip3 install pvcheetahdemo

  2. Run cheetah_demo_mic --access_key ${ACCESS_KEY}, replacing ${ACCESS_KEY} with your key.

Expected outcome

The Cheetah microphone demo starts for streaming speech-to-text transcription.

Guide limitations

  • Cheetah’s speech processing runs locally, but AccessKey validation requires internet connectivity.
  • The documented language support is English, French, German, Italian, Portuguese, and Spanish.
Documentation 1 ↗
Source-based practical guide

Build and run the C microphone demo

Clone the Cheetah repository, build its C demo, and start the microphone demo.

Before you start

  • A valid Picovoice AccessKey.
  • A suitable library path under the repository’s lib directory.
  • The default model file at lib/common/cheetah_params.pv, or a custom model file.
  • CMake and a C build environment.

Steps

  1. Clone the repository with submodules using SSH or HTTPS, as shown in the C demos section of the README.

  2. Build the demo with cmake -S demo/c/ -B demo/c/build && cmake --build demo/c/build.

  3. Run ./demo/c/build/cheetah_demo_mic -a ${ACCESS_KEY} -m ${MODEL_PATH} -l ${LIBRARY_PATH}, replacing each placeholder with its corresponding value.

Expected outcome

The C microphone demo runs with the supplied AccessKey, model, and library.

Guide limitations

  • AccessKey validation requires internet connectivity even though speech recognition runs offline.
  • The README does not specify platform-specific CMake or compiler prerequisites.
Documentation 1 ↗
Evidence and transparency

Sources & review

Official repository ↗
Source record details
Source availability
Active
Sources checked
13 October 2026
Original website title
Picovoice/cheetah · Official homepage title · retained in its source language
How we review sourcesSuggest a correction