BetterClaw
Models & Skills

LLM configuration

Choose the model that powers each agent and how it's authenticated.

LLM configuration

Every agent runs on a large language model that you choose. You configure the model per agent, so different agents in the same workspace can use different providers and models.

There are three ways to authenticate an agent's model — plus the option to run one on your own computer:

Bring your own key (BYOK)

Paste an API key from a provider like Anthropic, Google, or OpenAI. Required on Free.

ChatGPT subscription

Sign in with your existing ChatGPT Plus/Pro subscription instead of an API key.

Managed credits (paid plans)

On Starter, Pro, and Business, agents get a managed OpenRouter key minted for them — no key to paste. Starter includes up to $5 of managed LLM usage. New workspaces share one managed key and spending allowance across their agents. Adding an agent or resetting its LLM settings does not replenish that allowance.

Local model on your computer

Serve a model from Ollama, LM Studio, or llama.cpp on your own computer with BetterClaw Desktop.

Where to configure it

LLM settings live per agent, on the Model tab of the agent editor. Open Settings ▸ Agents from the left rail and select the agent to configure. What the tab shows depends on your plan.

Free plan: LLM Configuration

  1. In the LLM Configuration card, choose a provider, then a model, then add your API key (or connect a ChatGPT subscription).
  2. Save & Redeploy — the agent's service redeploys with the new settings.
You can change the model without re-entering your API key, as long as you stay on the same provider. Switching providers requires a fresh key.

the agent editor's Model tab showing provider, model, and API key fields

On a paid plan, the Model tab lists the agent's connected models — each with its own key — and assigns two roles:

  • Primary — runs every task.
  • Backup — used automatically if the primary is unavailable. Any further connected models stay available as lower-priority fallbacks.

To add one, click Connect a model, pick a provider, paste its API key (or sign in with ChatGPT), tick one or more of its models — or choose Use a model ID not listed here — and click Connect model. A key you've already saved for that provider is reused automatically.

Switch roles from the Primary and Backup pickers, or from a model's ⋯ menu (Make primary, Make backup, Remove as backup, Disconnect & forget key). Switching the primary restarts the agent and keeps every saved key, so you can switch back anytime. The gear icon on a row opens Edit Settings to replace its key, reconnect ChatGPT, or change the thinking level. An agent can have up to 10 connected backup models.

Backup models are a paid-plan feature, decided by the workspace's plan — that includes AppSumo-licensed workspaces. Free agents run on a single model.

Below the model settings, the Local LLM Models card lets you run the agent on a model served from your own computer — see Local models.

You can also set the model during onboarding when you create your first agent.

Resetting to default

On plans that include managed LLM keys, use Reset to Default to clear the provider, model, and key and fall back to the managed OpenRouter default (GLM 5.3 Flash). The agent service redeploys with the default settings. Free workspaces are bring-your-own-key only, so the button isn't shown there.

See also