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 (3, 4, 5)

• iOS 16.0+

The Leopard iOS binding is available via Swift Package Manager or CocoaPods.

To import the package using SPM, open up your project's Package Dependencies in XCode and add:

https://github.com/Picovoice/leopard.git

To import it into your iOS project using CocoaPods, add the following line to your Podfile:

pod 'Leopard-iOS'

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.

Add the Leopard model file in Xcode:

1. Create a model in Picovoice Console or use a default model.
2. Add the model as a bundled resource by selecting Build Phases and adding it to Copy Bundle Resources step.

Create an instance of the engine:

import Leopard

let modelPath = Bundle(for: type(of: self)).path(
        forResource: "${MODEL_FILE}", // Name of the model file name for Leopard
        ofType: "pv")!

let accessKey = "${ACCESS_KEY}" // AccessKey obtained from https://console.picovoice.ai/access_key
let leopard = Leopard(accessKey: accessKey, modelPath: modelPath)

Alternatively, you can provide modelPath as an absolute path to the model file on device.

Transcribe an audio file either by passing the absolute path or an url to the file:

do {
    let audioPath = Bundle(for: type(of: self)).path(forResource: "${AUDIO_FILE_NAME}", ofType: "${AUDIO_FILE_EXTENSION}")
    var result = leopard.process_file(audioPath)
    print(result.transcript)

    let audioURL = Bundle(for: type(of: self)).url(forResource: "${AUDIO_FILE_NAME}", withExtension: "${AUDIO_FILE_EXTENSION}")
    result = leopard.process_file(audioURL)
    print(result.transcript)
} catch let error as LeopardError {
    // handle error
} catch { }

Replace ${ACCESS_KEY} with yours obtained from Picovoice Console, ${MODEL_FILE} with the name of the Leopard model file name, ${AUDIO_FILE_NAME} with the name of the audio file and ${AUDIO_FILE_EXTENSION} with the extension of the audio file. Finally, when done be sure to explicitly release the resources using leopard.delete().

Default models for supported languages can be found in lib/common.

Create custom language models using the Picovoice Console. Here you can train language models with custom vocabulary and boost words in the existing vocabulary.

Pass in the .pv file via the modelURL or modelPath constructor argument:

let leopard = Leopard(accessKey: accessKey, modelPath: "${MODEL_FILE_PATH")
// or
let leopard = Leopard(accessKey: accessKey, modelURL: "${MODEL_FILE_URL}")

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.

Copy your AccessKey into the accessKey variable in LeopardAppTestUITests.swift. Open LeopardAppTest.xcodeproj with XCode and run the tests with Product > Test.

For example usage refer to our iOS demo application.