Back to Guides

Configure AI Models

Connect your preferred AI model providers and manage your model library

Estimated time: 5-10 minutes

Prerequisites

  • SetupClaw APP installed and environment check completed
  • An API key from at least one AI model provider
1

Step 1: Open Models & API

Navigate to the model configuration page where you can manage all your AI providers.

  1. aOpen SetupClaw APP and go to the Dashboard
  2. bClick "Models & API" in the sidebar
  3. cYou'll see the full provider catalogue and any models you've already configured
2

Step 2: Choose a Provider

Browse 30+ supported AI model providers. Major providers include OpenAI (GPT-4, GPT-4o), Anthropic (Claude 3.5, Claude 4), Google (Gemini 2.0, Gemini 2.5), DeepSeek, Mistral, Zhipu, Baichuan, Moonshot, Yi, Qwen, and any OpenAI-compatible custom endpoint.

  1. aScroll through the provider list or use the search bar to find your provider
  2. bClick on a provider card to expand its configuration panel
  3. cReview the available models for that provider
  4. dFor custom endpoints (any OpenAI-compatible API), select "Custom Provider" at the bottom of the list
3

Step 3: Enter API Key

Authenticate with your chosen provider by entering your API key. Some providers support one-click OAuth for a seamless experience.

  1. aClick the provider you want to configure
  2. bEnter your API key in the credential field. For providers like Google Gemini that support OAuth, click "Authorise with OAuth" — no API key needed
  3. cClick "Save" to store your credentials securely
4

Step 4: Set Default Model

Choose a default model that will be used for all chat conversations unless a specific agent overrides it.

  1. aIn the Models & API page, locate the "Default Model" section at the top
  2. bSelect your preferred model from the dropdown — this is used when no agent-specific model is configured
  3. cClick "Save" to confirm your default
5

Step 5: Configure Fallbacks

Set up a fallback chain so your AI assistant stays available even if your primary model goes down. If one model is unavailable, the system automatically tries the next one.

  1. aOpen the "Fallback Chain" settings below the default model selector
  2. bAdd models in priority order. Example: Claude → GPT-4 → Gemini
  3. cClick "Save" — the system will automatically switch to the next model if the current one is unavailable
6

Step 6: Verify Connection

Test that your API keys are valid and the models are reachable before you start chatting.

  1. aClick the "Verify" button next to each configured provider
  2. bThe dashboard will show a green checkmark for successful connections and a red indicator for failures
  3. cIf verification fails, double-check your API key and ensure your provider account has available credits

Tips

  • You can use models from multiple providers simultaneously — mix and match to find what works best for your use case
  • Different agents can use different models. Configure model preferences per agent in the Agents page
  • OAuth-supported providers (like Google Gemini) don't require manually copying API keys — just click to authorise

Troubleshooting

Back to Guides