Leopard Binding for Flutter

Leopard Speech-to-Text Engine

Made in Vancouver, Canada by Picovoice

Leopard is an on-device speech-to-text engine. Leopard is:

• Private; All voice processing runs locally.
• Accurate
• Compact and Computationally-Efficient
• Cross-Platform:
  • Linux (x86_64), macOS (x86_64, arm64), Windows (x86_64)
  • Android and iOS
  • Chrome, Safari, Firefox, and Edge
  • Raspberry Pi (4, 3) and NVIDIA Jetson Nano

Compatibility

This binding is for running Leopard on Flutter 2.8.1+ on the following platforms:

• Android 5.0+ (API 21+)
• iOS 13.0+

Installation

To start, you must have the Flutter SDK installed on your system. Once installed, you can run flutter doctor to determine any other missing requirements.

To add the Leopard plugin to your app project, you can reference it in your pub.yaml:

dependencies:
  leopard_flutter: ^<version>

AccessKey

Leopard requires a valid Picovoice AccessKey at initialization. AccessKey acts as your credentials when using Leopard SDKs. You can get your AccessKey for free. Make sure to keep your AccessKey secret. Signup or Login to Picovoice Console to get your AccessKey.

Permissions

To enable recording with the hardware's microphone, you must first ensure that you have enabled the proper permissions on both iOS and Android.

On iOS, open your Info.plist and add the following line:

<key>NSMicrophoneUsageDescription</key>
<string>[Permission explanation]</string>

On Android, open your AndroidManifest.xml and add the following line:

<uses-permission android:name="android.permission.RECORD_AUDIO" />
<uses-permission android:name="android.permission.INTERNET" />

Usage

An instance of Leopard is created by passing a model file path into its static constructor create:

import 'package:leopard_flutter/leopard.dart';

String accessKey = '{ACCESS_KEY}' // AccessKey obtained from Picovoice Console (https://console.picovoice.ai/)
String modelPath = '{MODEL_FILE_PATH}' // path relative to the assets folder or absolute path to file on device

void createLeopard() async {
    try {
        _leopard = await Leopard.create(accessKey, modelPath);
    } on LeopardException catch (err) {
        // handle Leopard init error
    }
}

Transcribe an audio file by passing in the path:

String audioPath = '{AUDIO_FILE_PATH}'

try {
    LeopardTranscript result = await _leopard.processFile(audioPath);
    print(result.transcript);
} on LeopardException catch (err) { }

When done, resources must be released explicitly:

await _leopard.delete();

Language Model

Add the Leopard model file to your Flutter application:

1. Create a model in Picovoice Console or use one of the default language models.
2. Add the model file to an assets folder in your project directory.
3. Add the asset to your pubspec.yaml:

flutter:
  assets:
    - assets/leopard_model.pv

4. In this example, the path to the model file in code would then be as follows:

String modelPath = "assets/leopard_model.pv";

Alternatively, if the model file is deployed to the device with a different method, the absolute path to the file on device can be used.

Word Metadata

Along with the transcript, Leopard returns metadata for each transcribed word. Available metadata items are:

• Start Time: Indicates when the word started in the transcribed audio. Value is in seconds.
• End Time: Indicates when the word ended in the transcribed audio. Value is in seconds.
• Confidence: Leopard's confidence that the transcribed word is accurate. It is a number within [0, 1].
• Speaker Tag: If speaker diarization is enabled on initialization, the speaker tag is a non-negative integer identifying unique speakers, with 0 reserved for unknown speakers. If speaker diarization is not enabled, the value will always be -1.

Demo App

For example usage, refer to our Flutter demo application.