Overview
Each AI provider in Whim has a different authentication method. CCR + OpenRouter works out of the box, Claude requires an OAuth token from your Anthropic subscription, and Codex uses a ChatGPT device code login.CCR + OpenRouter Setup
No setup required. CCR + OpenRouter is the default provider and is managed entirely by Whim. API usage is tracked and billed as part of your CU allocation. When you create a new workspace, CCR + OpenRouter is automatically selected and ready to use.Claude Subscription Setup
Connecting your Anthropic Claude subscription lets you run the native Claude Code CLI in Whim task containers.Getting Your OAuth Token
Open Claude Code on your local machine
If you don’t have Claude Code installed locally, follow Anthropic’s setup guide to install and authenticate.
Copy your OAuth token
Run the following command in your local terminal:This outputs your OAuth token. Copy it to your clipboard.
Token Management
- Status display — The Claude settings panel shows when your token was created and last used
- Updating — Paste a new token to replace the existing one
- Removing — Clear the token field and save to disconnect
Codex Subscription Setup
Connecting your ChatGPT account lets you run the native Codex CLI in Whim task containers.Device Code Login Flow
Enter the device code
Whim displays a device code (e.g.,
AB1D-EF2H) and a link to the OpenAI verification page. Click the link to open it in your browser.Authorize the connection
Enter the device code on the OpenAI verification page and approve the connection. The page will confirm when authorization is complete.
Device codes expire after a short time. If the code expires before you complete authorization, click Sign in to ChatGPT again to generate a new one.
Setting Workspace Defaults
Workspace defaults determine the provider and model used when members create new tasks. Only workspace admins can change these. Go to Settings > Workspace > Defaults to configure:- Default provider — Which provider new tasks use (CCR + OpenRouter, Claude, or Codex)
- Default model — Which model new tasks use within that provider
- Permission mode — How the agent handles tool permissions
- Reasoning effort — Codex reasoning level (see Fast Mode & Reasoning)
- Fast mode — Whether fast mode is enabled by default
Provider-Specific Settings
Each provider supports its own set of configurable options:| Setting | CCR + OpenRouter | Claude | Codex |
|---|---|---|---|
| Model | Yes | Yes | Yes |
| Permission mode | Yes | Yes | Yes |
| Fast mode | No | Yes | Yes (GPT 5.4 only) |
| Reasoning effort | No | No | Yes |
Connection Status Indicators
The settings panel shows the connection state for each provider:| Indicator | Meaning |
|---|---|
| Green dot — Connected | Provider is authenticated and ready to use |
| Grey dot — Not connected | Provider needs setup before it can be used |
| Spinner — Checking | Whim is verifying the connection status |
Troubleshooting
Claude token shows 'Not connected' after saving
Claude token shows 'Not connected' after saving
Your OAuth token may have expired. Generate a fresh token by running
claude setup-token on your local machine and paste the new token in Whim.Codex device code expired before I could enter it
Codex device code expired before I could enter it
Device codes are time-limited. Click Sign in to ChatGPT again to generate a new code. Make sure to complete the authorization promptly after the code appears.
Tasks fail immediately after switching providers
Tasks fail immediately after switching providers
Make sure the provider is connected (green dot) before setting it as your default. If you’ve just connected, try creating a new task — existing tasks keep the provider they were started with.
My user defaults aren't taking effect
My user defaults aren't taking effect
User defaults override workspace defaults, but only for new tasks. Check that you’re looking at Settings > My Defaults (not workspace defaults). Also verify that the provider you selected is connected.
Next Steps
Supported Providers
Compare providers and available models.
Fast Mode & Reasoning
Configure speed vs quality tradeoffs.

