Revoize SDK C API Reference

CModelType Enum

The CModelType enum defines the AI models available in the Revoize SDK:

typedef enum {
    /// **Capella:** A lightweight model designed for real-time applications.
    REVOIZE_MODEL_CAPELLA = 0,
    /// **Algol:** A generative model with advanced capabilities.
    REVOIZE_MODEL_ALGOL = 1,
    /// **Regulus:** A lightweight model designed for real-time applications.
    REVOIZE_MODEL_REGULUS = 2,
    /// **Octantis:** A lightweight model designed for real-time applications.
    REVOIZE_MODEL_OCTANTIS = 3
} CModelType;

revoize_init Function

Signature

int revoize_init(CModelType model_type);

The revoize_init function is your starting point for using the Revoize SDK. It sets up the SDK with the specified model type. Note: Make sure to call this function before any other SDK functions to ensure everything is properly set up.

Parameters

• model_type The model type to initialize the SDK with (0=Capella, 1=Algol, 2=Regulus, 3=Octantis).

Returns

• 0 if initialization is successful.
• Non-zero if an error occurs during initialization (e.g., session creation, state setup, or missing model constants).

Example

#include "revoize_sdk.h"

int main() {
    // Initialize the SDK with the Capella model.
    int result = revoize_init(REVOIZE_MODEL_CAPELLA);
    if (result != 0) {
        fprintf(stderr, "Failed to initialize Revoize SDK\n");
        return 1;
    }
    return 0;
}

revoize_process Function

Signature

int revoize_process(
    const float* input,
    size_t input_len,
    float* output,
    size_t* output_len
);

The revoize_process function takes an input audio chunk and returns the enhanced chunk.

WARNING

• Ensure revoize_init() is called before using this function.
• The input should be from a 48 kHz recording.
• The output buffer must have enough space to store the enhanced audio samples.

Parameters

• input Pointer to input audio samples (48 kHz, single channel).
• input_len Number of input samples.
• output Pointer to output buffer where enhanced audio will be written.
• output_len Pointer to store the number of output samples.

Returns

• 0 if processing is successful.
• Non-zero if an error occurs during processing (e.g., if the SDK has not been initialized).

Safety

This function is unsafe because it takes raw pointers. The caller must ensure:

• input points to a valid array of input_len samples
• output points to a valid array with enough space for the output
• output_len points to a valid integer

Example

#include "revoize_sdk.h"
#include <stdio.h>
#include <stdlib.h>

int main() {
    // Initialize the SDK before processing audio.
    int result = revoize_init(REVOIZE_MODEL_CAPELLA);
    if (result != 0) {
        fprintf(stderr, "Failed to initialize Revoize SDK\n");
        return 1;
    }

    // Prepare an audio chunk from a 48 kHz recording.
    const size_t num_samples = 480 * 10; // 10 frames of 480 samples each
    float* input = (float*)malloc(num_samples * sizeof(float));
    if (input == NULL) {
        fprintf(stderr, "Failed to allocate input buffer\n");
        return 1;
    }

    // Fill input with some test data
    for (size_t i = 0; i < num_samples; i++) {
        input[i] = 1.0f;
    }

    // Allocate output buffer
    float* output = (float*)malloc(num_samples * sizeof(float));
    if (output == NULL) {
        free(input);
        fprintf(stderr, "Failed to allocate output buffer\n");
        return 1;
    }

    // Process the audio
    size_t output_len = num_samples;
    result = revoize_process(input, num_samples, output, &output_len);
    if (result != 0) {
        free(input);
        free(output);
        fprintf(stderr, "Failed to process audio\n");
        return 1;
    }

    printf("Enhanced audio has %zu samples\n", output_len);

    // Clean up
    free(input);
    free(output);
    return 0;
}