Providers
groq
Groq — ultra-low-latency inference on custom LPU hardware. Llama + Mixtral + Gemma.
import { groq } from '@agentskit/adapters'
const adapter = groq({
apiKey: process.env.GROQ_API_KEY!,
model: 'llama-3.3-70b-versatile',
})#Options
| Option | Type | Default |
|---|---|---|
apiKey | string | required |
model | string | llama-3.3-70b-versatile |
baseUrl | string | https://api.groq.com/openai/v1 |
retry | RetryOptions | inherited |
#Capabilities
{ streaming: true, tools: true, usage: true } — Groq exposes a strict OpenAI-compatible surface, so the request shape matches openai({ baseUrl }).
#Why groq
- Sub-100 ms first-token latency — best for realtime voice + chat.
- OpenAI-compatible.
#Env
| Var | Purpose |
|---|---|
GROQ_API_KEY | API key |
#Related
Explore nearby
- PeerProviders
25 native chat and embedder adapters, plus higher-order adapters that compose candidates. Separate from the 140-provider models catalog.
- PeerChoosing an adapter
Capability decision table and rules of thumb for picking a chat adapter.
- PeerHosted chat adapters
17 managed-LLM adapters. Same contract; swap by changing one import.