Cheetah Speech-to-Text
C Quick Start

Platforms

• Linux (x86_64)
• macOS (x86_64, arm64)
• Windows (x86_64)
• NVIDIA Jetson Nano
• Raspberry Pi (3, 4)

Requirements

• C99-compatible compiler
• CMake (3.13+)
• For Windows Only: MinGW is required to build the demo

Picovoice Account & AccessKey

Signup or Login to Picovoice Console to get your AccessKey. Make sure to keep your AccessKey secret.

Quick Start

Setup

1. Clone the repository:

git clone --recurse-submodules https://github.com/Picovoice/cheetah.git

Usage

1. Include the public header files (picovoice.h and pv_cheetah.h).
2. Link the project to an appropriate precompiled library for the target platform and load it.
3. Download a custom model from Picovoice Console or the default model.
4. Construct the Cheetah object:

static const char* ACCESS_KEY = "${ACCESS_KEY}";

const char *model_file_path = "${MODEL_FILE_PATH}";

float endpoint_duration_sec = 1.f;

bool enable_automatic_punctuation = false;

pv_cheetah_t *cheetah;

const pv_status_t status = pv_cheetah_init(
    ACCESS_KEY
    model_file_path,
    endpoint_duration_sec,
    enable_automatic_punctuation,
    &cheetah);

if (status != PV_STATUS_SUCCESS) {
    // error handling logic
}

6. Pass in an audio path to the pv_cheetah_process function:

extern const int16_t *get_next_audio_frame(void);

while (true) {
   const int16_t *pcm = get_next_audio_frame();

   char *partial_transcript = NULL;
   bool is_endpoint = false;
   const pv_status_t status = pv_cheetah_process(cheetah, pcm, &partial_transcript, &is_endpoint);
   if (status != PV_STATUS_SUCCESS) {
     // error handling logic
   }
   fprintf(stdout, "%s", partial_transcript);
   free(partial_transcript);

   if (is_endpoint) {
      char *final_transcript = NULL;
      status = pv_cheetah_flush(cheetah, &final_transcript);
      if (status != PV_STATUS_SUCCESS) {
          // error handling logic
      }
      fprintf(stdout, "%s\n", final_transcript);
      free(final_transcript);
   }
}

8. Release resources explicitly when done with Cheetah:

pv_cheetah_delete(cheetah);

Demo

For the Cheetah SDK, we offer demo applications that demonstrate how to use the Speech-to-Text engine on real-time audio streams (i.e. microphone input) and audio files.

Setup

1. Clone the Cheetah repository from GitHub using HTTPS:

git clone --recurse-submodules https://github.com/Picovoice/cheetah.git

2. Build the microphone demo:

cd cheetah
cmake -S demo/c/. -B demo/c/build
cmake --build demo/c/build --target cheetah_demo_mic

Usage

To see the usage options for the demo:

./demo/c/build/cheetah_demo_mic

Run the command corresponding to your platform from the root of the repository:

./demo/c/build/cheetah_demo_mic \
-a ${ACCESS_KEY} \
-m lib/common/cheetah_params.pv \
-l lib/${PLATFORM}/${ARCH}/libpv_cheetah.so \
-d ${AUDIO_DEVICE_INDEX}

For more information on our Cheetah demos for C, head over to our GitHub repository.

Resources