Groq Provider Setup: Auth, Model Selection, and Whisper Transcription

Groq delivers extremely fast inference across open-weight models (Llama, Gemma, Kimi, Qwen, GPT OSS, and others) by relying on custom LPU hardware. Through the Groq plugin, two providers are registered: one chat provider compatible with OpenAI and one audio media-understanding provider.

Install plugin

After installing the official plugin, restart Gateway:

openclaw plugins install @openclaw/groq-provider
openclaw gateway restart

Getting started

Get an API key

Generate an API key at console.groq.com/keys.

Set the API key

export GROQ_API_KEY=gsk_...

Set a default model

{
  agents: {
    defaults: {
      model: { primary: "groq/llama-3.3-70b-versatile" },
    },
  },
}

Verify the catalog is reachable

openclaw models list --provider groq

Config file example

{
  env: { GROQ_API_KEY: "gsk_..." },
  agents: {
    defaults: {
      model: { primary: "groq/llama-3.3-70b-versatile" },
    },
  },
}

Built-in catalog

OpenClaw includes a manifest-based Groq catalog containing both reasoning and non-reasoning models. Execute openclaw models list --provider groq to display the static entries for your current version, or visit console.groq.com/docs/models for Groq's full list.

Tip

The catalog changes with each OpenClaw release. Use openclaw models list --provider groq to see the rows recognized by your installed version; verify against console.groq.com/docs/models for models that are new or deprecated.

Reasoning models

For Groq reasoning models (marked with reasoning: true in the table), OpenClaw maps its shared /think levels to reasoning_effort values of low, medium, or high. When /think off or /think none is used, reasoning_effort is left out of the request entirely instead of being sent with a disabled value.

Refer to Thinking modes for the shared /think levels and how OpenClaw translates them for each provider.

Audio transcription

The Groq plugin also registers an audio media-understanding provider, so voice messages can be transcribed through the shared tools.media.audio interface.

To configure Groq as the default audio backend:

{
  tools: {
    media: {
      audio: {
        models: [{ provider: "groq" }],
      },
    },
  },
}

Environment availability for the daemon

When Gateway runs as a managed service (launchd, systemd, Docker), GROQ_API_KEY must be accessible to that process, not only to your interactive shell.

Warning

A key exported solely in an interactive shell will not be available to a launchd or systemd daemon unless that environment is also imported there. Set the key in ~/.openclaw/.env or through env.shellEnv so the gateway process can read it.

Custom Groq model ids

OpenClaw accepts any Groq model identifier at runtime. Provide the exact id that Groq displays and prepend groq/. The static catalog handles the most frequent cases; ids not in the catalog fall back to the default OpenAI-compatible template.

{
  agents: {
    defaults: {
      model: { primary: "groq/<your-model-id>" },
    },
  },
}

Related

• Model providers, Picking providers, model references, and how failover works.
• Thinking modes, Reasoning intensity levels and how they interact with provider policies.
• Configuration reference, Complete configuration schema, covering provider and audio options.
• Groq Console, The Groq dashboard, API documentation, and pricing information.