DF Radio is a Laravel-based speech-processing prototype. It accepts an audio URL and language, downloads the source file, and uses Google Cloud Speech-to-Text to return a transcript, confidence data, and word-level timing information.
Work in progress. The current repository contains the speech-processing experiment alongside Laravel authentication and the default application shell.
- Audio transcription for MP3 and FLAC input.
- Language selection for the recognition request.
- Transcript confidence and word timing data.
- A Blade-based interface for submitting audio and viewing results.
- Laravel authentication scaffolding and standard web/API routes.
- Laravel 8 and PHP
- Blade templates
- Vue 2 with Laravel Mix
- Google Cloud Speech-to-Text client
- MySQL-compatible Laravel database configuration
- PHP 7.3 or newer
- Composer
- Node.js and npm
- A configured database
- Google Cloud Speech credentials
composer install
cp .env.example .env
php artisan key:generate
php artisan storage:link
npm installConfigure the database values in .env, then run migrations if your local setup needs the included user tables:
php artisan migrateThe current audio controller expects a Google Cloud service-account JSON file at storage/app/public/auth.json. Keep credentials outside version control and never commit that file.
Start the Laravel application:
php artisan serveIn a second terminal, compile the frontend assets:
npm run dev- Submit an audio URL and language through the web form.
- The application stores a local copy of the audio file.
- Google Cloud Speech-to-Text processes MP3 or FLAC input.
- The result is rendered with the transcript, confidence, and timing information.
The implementation currently lives in app/Http/Controllers/AudioController.php and the web routes in routes/web.php.
php artisan serve # Start the Laravel server
npm run dev # Watch and compile frontend assets
npm run watch # Watch assets using Laravel Mix
npm run production # Build production assetsDF Radio is an experimental speech-processing project. The repository is useful as a working prototype, but the audio pipeline, credential handling, and production deployment story still need hardening before public service use.
No open-source license has been declared yet. Until a license is added, reuse and redistribution are not granted by default.
Made with love by Laflamme.