Building Voice Assistants with Laravel: Speech-to-Text and Beyond

Voice as Interface

Voice interfaces are becoming ubiquitous. Building voice capabilities into your Laravel application opens new interaction paradigms—hands-free operation, accessibility improvements, and natural conversation.

Speech-to-Text with Whisper

class SpeechToText
{
    public function transcribe(string $audioPath): string
    {
        $response = Http::attach(
            'file',
            file_get_contents($audioPath),
            'audio.webm'
        )->post('https://api.openai.com/v1/audio/transcriptions', [
            'model' => 'whisper-1',
        ]);

        return $response->json('text');
    }
}

Text-to-Speech

class TextToSpeech
{
    public function synthesize(string $text, string $voice = 'alloy'): string
    {
        $response = Http::withHeaders([
            'Authorization' => 'Bearer ' . config('services.openai.key'),
        ])->post('https://api.openai.com/v1/audio/speech', [
            'model' => 'tts-1',
            'input' => $text,
            'voice' => $voice,
        ]);

        $path = 'audio/' . Str::uuid() . '.mp3';
        Storage::put($path, $response->body());

        return $path;
    }
}

Voice Conversation Loop

class VoiceAssistant
{
    public function processVoiceInput(string $audioPath): array
    {
        // Transcribe
        $text = $this->stt->transcribe($audioPath);

        // Process with chatbot
        $response = $this->chatbot->respond($text);

        // Synthesize response
        $audioResponse = $this->tts->synthesize($response);

        return [
            'transcription' => $text,
            'response_text' => $response,
            'response_audio' => $audioResponse,
        ];
    }
}

Real-Time Processing

class RealtimeVoice
{
    public function streamTranscription(Request $request): StreamedResponse
    {
        return response()->stream(function () use ($request) {
            $audioStream = $request->getContent();

            // Process in chunks
            foreach ($this->chunkAudio($audioStream) as $chunk) {
                $partial = $this->stt->transcribeChunk($chunk);
                echo "data: " . json_encode(['partial' => $partial]) . "\n\n";
                ob_flush();
                flush();
            }
        }, 200, ['Content-Type' => 'text/event-stream']);
    }
}

Conclusion

Voice interfaces add powerful capabilities to applications. Start with basic transcription and synthesis, then build toward real-time conversation. Consider accessibility implications and provide fallback text interfaces.