G
Groq
by Groq, Inc.
Run extremely fast chat completions and audio transcriptions on open-source models (Llama 3, Mixtral, Gemma) using Groq's LPU inference hardware.
AI ServicesFree / PaidAPI keyllmfast-inferencellamamixtralopen-sourceai
How agents use Groq
- ✓Agent uses Groq for low-latency real-time inference where response speed is critical
- ✓Agent transcribes audio recordings or voice memos using Whisper on Groq hardware
- ✓Agent uses Llama 3 70B for complex reasoning tasks at lower cost than frontier models
- ✓Agent runs structured JSON extraction with response_format: json_object for parsing unstructured data
- ✓Agent uses Groq as a fast fallback when primary LLM providers are throttled or unavailable
Agent actions
Chat CompletionGenerate a chat response using a Groq-hosted model with ultra-low latency.
Inputs: messages, model, systemPrompt, temperature, maxTokens, responseFormat
Returns: content, finishReason, promptTokens, completionTokens, model, tokensPerSecond
Transcribe AudioTranscribe audio to text using Whisper on Groq hardware.
Inputs: audioUrl, language, prompt
Returns: text, language, duration
List ModelsRetrieve all available models on Groq.
Returns: models, modelCount
Example workflows
1
Real-time chat agent
An agent that needs sub-second response times uses Groq with Llama 3 for low-latency interactive processing.
About Groq
- Vendor
- Groq, Inc.
- Pricing Always review details with the vendor
- Free / Paid — Generous free tier available. Paid usage billed per million tokens, typically cheaper than OpenAI equivalents.
- Authentication
- API key
- Rate limit Always review details with the vendor
- 30 requests / minute
- Compatible nodes
- AgentResource
- Website
- https://groq.com
Put Groq behind your AI agent
Set up Groq alongside your other tools in the planner — no code — so AI can act on it for your customers.
Open the planner