Skip to content

Setting Up Your AI Provider

Provider overview

Every OneClickClaw plan is bring-your-own-AI: your agent answers with the AI subscription or API key you connect. There is no built-in free model, so your agent answers nobody until you connect one. You connect it inside your agent itself. The OneClickClaw dashboard has no provider form; it points you to the right place.

You have three ways to connect:

  • Your ChatGPT plan. Needs ChatGPT Plus or higher. No separate API bill.
  • Your SuperGrok or X Premium+ plan. Sign in with your xAI account. No separate API bill.
  • An API key. From OpenAI, Anthropic, Google, OpenRouter and many more providers. Claude works with an Anthropic API key.

A Hermes Agent server can also sign in with a Nous Portal or MiniMax plan.

Server page - AI Model tab (OpenClaw)

Your AI model lives inside your agent

Short steps for connecting a provider inside your agent.

Open agent

Opens your agent so you can connect a provider.

Step-by-step guide

Opens the full guide for signing in with ChatGPT.

Step-by-step setup

On an OpenClaw server

1

Open your agent

On the dashboard overview, click Open Agent on your server card. While no model is connected, the overview also shows a "Setup Incomplete" banner with an Open agent to pick a model button.

2

Go to Settings > Models

Inside your agent, open Settings, then Models, and press Connect provider.

3

Pick how to connect

For a ChatGPT plan: first, in ChatGPT, turn on Settings > Security and login > App security > Enable device code sign-in. Then in your agent pick OpenAI > Codex login (device code) and follow the code it shows. Older servers may still show the old name "ChatGPT Device Pairing" for this button.

For Grok: pick xAI (Grok) > xAI OAuth. It needs SuperGrok or X Premium+.

For an API key: pick your provider and paste the key.

4

Choose your default model

After adding a provider, pick it in the Default model dropdown. OpenClaw checks the model before it turns conversations on.

The full walkthrough with pictures is in Connect ChatGPT.

On a Hermes Agent server

While your Hermes server is being built, the build page asks you to choose the AI behind your agent. If you pick a plan to sign in with and keep the build page open, the sign-in opens by itself when the server is ready. You can also do it later:

1

Open the AI Model tab

On the overview, click Control Panel on your Hermes server card, then open the AI Model tab. The overview strip "Your Hermes agent is live" also has a Connect your AI button that goes there.

2

Sign in or add a key

Under "Use a subscription you already have", pick Sign in with ChatGPT, Grok, Nous Portal or MiniMax. A code appears; approve it with your provider. Or choose an AI Provider and add an API key.

3

Wait for the check

After a sign-in, your server asks the model one short question before it tells you the AI is connected.

Tip

Already pay for ChatGPT Plus or higher, or for SuperGrok or X Premium+? Sign in with that plan. You need no API key and get no separate API bill.

Where to get API keys

If you connect with an API key, you first need one from your provider. Here is how to get one from the most popular providers. (If you sign in with a ChatGPT, SuperGrok or X Premium+ plan, you can skip this part.)

OpenAI

1

Create an OpenAI account

Visit platform.openai.com and sign up or log in.

2

Go to API Keys

Click your profile icon in the top-right corner, then select API keys from the dropdown menu.

3

Create a new secret key

Click Create new secret key, name it (for example, "OneClickClaw Agent"), and copy the key immediately. OpenAI only displays the full key once.

Anthropic

1

Create an Anthropic account

Visit console.anthropic.com and sign up or log in.

2

Navigate to API Keys

Open the Settings page and click API Keys.

3

Create a new key

Click Create Key, give it a descriptive name, and copy the key immediately. Anthropic only shows the full key once.

OpenRouter

1

Create an OpenRouter account

Visit openrouter.ai and sign up. OpenRouter supports Google, GitHub, and email sign-in.

2

Generate an API key

Navigate to Keys in your account settings and create a new key. OpenRouter uses credits that you top up as needed. Each model's price is listed on its OpenRouter page.

Google (Gemini)

1

Open Google AI Studio

Visit aistudio.google.com and sign in with your Google account.

2

Create an API key

Click Get API key in the left sidebar. Create a new key or use an existing one.

Note

A free Google AI Studio key can be refused on the newest Gemini models. On OpenClaw, if your agent reports a quota or billing error with a free key, pick a lighter Gemini model in the Default model dropdown.

xAI (Grok)

1

Create an xAI account

Visit console.x.ai and sign up or log in.

2

Create an API key

Navigate to the API Keys section and generate a new key. Copy it and store it securely.

ProviderWebsiteBest For
OpenAIplatform.openai.comGeneral purpose, vision, most widely supported
Anthropicconsole.anthropic.comReasoning, safety, long context windows
OpenRouteropenrouter.aiModel exploration, many providers with one key
Googleaistudio.google.comLong context, research
xAIconsole.x.aiReal-time knowledge, conversational style

Warning

Never share your API key publicly or commit it to a code repository. Treat it like a password. If you suspect a key has been compromised, revoke it immediately from your provider's dashboard and generate a new one. For more on how your key is protected, see API Key Security.

Switching providers

You can change your AI provider at any time, inside your agent:

  • OpenClaw: open your agent, go to Settings > Models, press Connect provider to add another provider, then pick the model you want in the Default model dropdown. OpenClaw can hold several providers at once.
  • Hermes Agent: open the AI Model tab on the server page, sign in with another plan or pick another AI Provider, then use Change Model and Apply Model.
  • Your agent's personality files and your channel connections stay as they are. Only the AI behind the agent changes.

Note

Switching the AI needs no server restart.

API key security

Here is how your key is protected on your dedicated server:

  • Never logged. API keys are never written to application logs, error reports, or monitoring data. Our support team cannot see your key.
  • Stored on your server. A key you add inside your agent lives on your dedicated VPS and is destroyed along with the server when you delete your instance. The one exception, a key our ClawCrew support assistant replaces at your request, is described on the security page.
  • File permissions locked down. The configuration file containing your key has restricted permissions (chmod 600), readable only by the service account running your agent.

For a comprehensive overview of our security practices, see the API Key Security article. For information on BYOK and why you control your own key, see BYOK (Bring Your Own Key).