Voice Assistant Loop
A 3-step recipe with NVIDIA NIM. We can help you get access.
Recipe · 3 steps
Voice Assistant Loop
with NVIDIA NIM · Artificial Intelligence Management
Speech-to-text, chat answer, then text-to-speech in a single loop.
- Transcribe the uploaded audio clip into text using a Riva ASR NIM via a multipart/form-data upload.
- Send the transcript to a chat model to generate a spoken-style reply.
- Synthesize the chat reply back into audio bytes using a Riva TTS NIM.
What you'll need to have ready
- apiKey: NVIDIA developer API key (nvapi-...) sent as a Bearer token.
- audioFile: WAV/FLAC/MP3 audio clip to transcribe.
- asrModel: Riva ASR model id.
- chatModel: LLM model id used to answer the transcript.
- ttsModel: Riva TTS model id.
- voice: TTS voice identifier.
For your developer
Arazzo workflow nvidia-nim:voice-assistant-loop, 3 operations: createTranscription → createChatCompletion → createSpeech. Specs on APIs.io.