Feature/spech to text

[mappweb]

//Example
class SpeechToTextAgent extends Transcribe
{
public function transcribe(): TranscribeProviderInterface
{
return new OpenAITranscribeProvider(
config('neuron-ai.providers.openai.api_key'),
'whisper-1',
'es',
);
}
}

//Agent
$fullPath = '.../.../.oga'
$agent = SpeechToTextAgent::make();
$string = $agent->transcribeAudio(new AudioContent($fullPath, SourceType::URL));

[ilvalerione]

Ho @mappweb any PR should bring a comprehensive description of what you worked on, with code examples, use case, and unit tests.

There is also a template for publishing PRs that can guide you in providing all the necessary information.

[ilvalerione]

Furthermore you target the 2.x branch not main.

[mappweb]

Hi @ilvalerione. Added example.

[ilvalerione]

I don't understand how it integrates with the framework. It's seems a standalone implementation.

By the way, can you use appropriate code formatting? As I said it should also be documented, target the 2.x branch, and tested eventually.

[mappweb]

@ilvalerione An agent that interacts by voice. This would be the first layer of transcription.

[devinkel]

Any news about this feature?

[izdrail]

So @mappweb Why don't you keep that logic to own agent ?