Menú
Chunked sliding-window streaming speech-to-text via OpenAI Whisper HTTP API
Instalar con Codex o Claude Copia este prompt, pégalo en Codex, Claude u otro asistente, y deja que revise la página de la skill y la instale por ti.
Basado en la clasificación ocupacional SOC
Edit, transform, and enhance images using AI models
Extract text from images using OCR and vision AI
Neural text-to-speech via Amazon Polly
Speaker diarization — identifies and tracks who is speaking at each moment in an audio stream
Semantic endpoint detection — uses an LLM to classify whether the user's utterance is a complete thought, reducing false turn boundaries on mid-sentence pauses
Batch speech-to-text via Google Cloud Speech-to-Text API
| name | streaming-stt-whisper |
| description | Chunked sliding-window streaming speech-to-text via OpenAI Whisper HTTP API |
| category | voice |
Streaming speech-to-text using OpenAI's Whisper model via the /v1/audio/transcriptions HTTP API.
Audio is accumulated in a sliding-window ring buffer and sent as 1-second WAV chunks with 200 ms
overlap for continuity.
Set OPENAI_API_KEY in your environment or agent secrets store.
ws, no native binaries)prompt for cross-chunk continuityspeech_start / speech_end eventserror and continues processing — no session crashIn agent.config.json:
{
"voice": {
"stt": "whisper"
}
}
Provider-specific options via providerOptions:
{
"voice": {
"stt": "whisper",
"providerOptions": {
"model": "whisper-1",
"language": "en",
"baseUrl": "https://api.openai.com"
}
}
}
| Event | Payload | Description |
|---|---|---|
interim_transcript | TranscriptEvent | Emitted after each chunk is transcribed |
final_transcript | TranscriptEvent | Emitted after flush() completes |
speech_start | — | RMS energy crossed threshold (0.01) |
speech_end | — | RMS energy dropped below threshold |
error | Error | Fetch failure (session continues) |
close | — | Session fully terminated |