← Back

ofxWhisper

github.com/tettou771/ofxWhisper · @tettou771

OF 0.12 Audio Artificial Intelligence

Transcribes recorded or live microphone audio in real time using the OpenAI Whisper ASR API.

5 0 Last commit 3y ago Created 3y ago

Branches ahead of main

These branches have commits not yet on main — could be unmerged fixes or work in progress worth a look.

windowsSoundBug 1 commit ahead · last commit 3y ago

README

ofxWhisper

ofxWhisper is an openFrameworks addon for transcribing recorded audio or audio files in real-time. This addon sends audio data and uses the OpenAI Whisper ASR API to convert the audio data into text.

Features

Transcribe audio file
Real-time transcription of mic input
Thread-safe implementation

Usage

Setup

1. Copy the ofxWhisper folder to the of_v0.X.X_addons/ directory. 2. Add the following dependencies to your project:

Create and initialize an instance

#include "ofxWhisper.h"

ofxWhisper whisper;

void ofApp::setup() {
    whisper.setup("your_api_key_here");
    whisper.setupRecorder(); // Optionally pass a sound device ID
}

Transcribe from file

     whisper.transcript("sample_speech.mp3");



 void ofApp::update() {     // Check if the transcript is ready     if (whisper.hasTranscript()) {         transcript = whisper.getNextTranscript();         transcriptReady = true;     } }
Record and transcript
 // Space key to record and transcript void ofApp::keyPressed(int key){ 	if(key == ' '){ 		if(recorder.isRecording()){ 			recorder.stopRecording(); 		}else{ 			recorder.startRecording(ofToDataPath(ofGetTimestampString()+".wav", true)); 		} 	} }

Get this addon

HTTPS

SSH

View on GitHub · Open with Github Desktop · Download ZIP