Intro to AI Gateway

AI Gateway lets your app call AI models from an API Endpoint. You can generate text, generate images, or create embeddings, then use the result in the rest of the workflow or send it back to the interface.

You do not need to install an integration. The actions appear under the AI category when you add an action in a backend workflow.

These actions run on the server, so prompts and API keys stay private.

In the AI subtab of Data & API, you can also add provider keys and check usage.

THIS IS NOT WEWEB AI

WeWeb AI helps you build the app in the editor. AI Gateway is for your published app, for example summarizing a support message or generating a product image.

Before you start

AI Gateway needs the WeWeb backend. If you do not see the AI tab in Data & API, set up the backend first.

When you add an AI action, use Provider on the right to choose Google Gemini, OpenAI, or Anthropic.

Google Gemini can run with no API key. If you choose OpenAI or Anthropic without a key, the action will not run until you add one.

Learn how to choose a provider →

Add an AI action

1. Go to the Data & API tab, then open the Backend Workflows subtab.
2. Open API Endpoints and create or open an endpoint.
3. Click Add an action.
4. Open the AI category and choose Text Generation, Image Generation, or Embed Text.
5. Choose a Provider and Model, then fill in the prompt or other fields.
6. Use Send Response to return the result, or pass it to the next action.
7. Call the endpoint from an interface workflow.

Learn how to call an API Endpoint from the interface →

What you can do

• Providers: choose Google Gemini, OpenAI, or Anthropic, and add an API key when you need one.
• Usage: see recent AI calls and how much of the free $1 credit is left.
• AI Actions: pick Text Generation, Image Generation, or Embed Text for your use case.

Related pages

• Intro to API Endpoints →
• Call API Endpoints from the interface →
• Handling streaming responses →