Yes — you can use Groq models instead of Google or OpenAI for STT, TTS, and LLM tasks in LiveKit Agents. What you can do: • Use Groq STT models (e.g.,

whisper-large-v3

,

whisper-large-v3-turbo

) for transcription. • Use Groq TTS (e.g.,

playai-tts

, voices like

Arista-PlayAI

) for speech synthesis. • Use Groq LLMs (e.g.,

gpt-oss-120b

,

llama3-8b-8192

) via the Groq plugin or the OpenAI-compatible integration. Basic setup steps: 1. Install the Groq-enabled package for your runtime (examples:

uv add "livekit-agents[groq]~=1.4"

for Python, or the OpenAI-compatible plugin for Node). 2. Set your Groq API key in your environment (e.g.,

GROQ_API_KEY

). 3. Specify the Groq model when creating the LLM/STT/TTS client (examples:

llm="groq/gpt-oss-120b"

or

groq.LLM(model="llama3-8b-8192")

). Would you like a short example for Python or Node.js showing how to configure a Groq LLM or STT/TTS? Sources: Groq and LiveKit | LiveKit Documentation | Groq STT plugin guide | LiveKit Documentation | Groq TTS plugin guide | LiveKit Documentation | Groq LLM plugin guide | LiveKit Documentation | OpenAI-compatible LLMs | LiveKit Documentation, with interactive elements