Gemini: Google AI for text, images & embeddings

Call Google Gemini from your WeWeb workflows

WeWeb has a native Gemini integration through its AI Gateway. It lets your published app call Google's Gemini models directly from backend workflows, with no plugin to install and no API key required to get started.

With the AI Gateway, you can:

Call Gemini for text generation inside a backend workflow, no integration install needed
Generate images from a text prompt and show them in the app or upload them to storage
Turn text into vector embeddings for semantic search or grouping similar records
Run Gemini through WeWeb Cloud with no Google API key, using the workspace's free $1 credit
Swap in your own Google API key per environment (Editor, Staging, Production) to bill Gemini calls to your own account
Return structured output (an object, a list, or one option from a set of choices) instead of free text
Stream a Gemini reply so the UI shows words as they arrive
Track every Gemini call's model, cost, and originating workflow under AI Usage

Gemini is Google's family of multimodal AI models for text, image, and code generation. WeWeb is a no-code web app builder that generates a standard Vue.js application you can export and self-host; its AI Gateway is the part of WeWeb's native backend (Data & API) that calls AI providers like Gemini on the server, so your app never has to talk to Google's API directly.

What WeWeb supports natively

WeWeb's AI Gateway covers three Gemini-powered actions (text generation, image generation, and embeddings), zero-key setup through WeWeb Cloud, per-environment API keys, and usage tracking, all from Data & API → AI.

Category	Feature	What it does
Plugin setup	AI Gateway (Data & API → AI)	Adds Google Gemini, OpenAI, and Anthropic as selectable providers for AI actions, no plugin install required.
Provider setup	Google Gemini as default provider	Preselects Gemini when you add a new AI action, ahead of OpenAI and Anthropic.
Credentials	WeWeb Cloud for Gemini	Runs Gemini calls with no Google API key, using the workspace's free $1 credit and workspace billing after that.
Credentials	Custom Gemini API key	Lets you paste your own Google API key under Data & API → AI → AI Providers, per environment, so usage bills to your Google account instead of WeWeb Cloud.
Actions	Text Generation action	Asks a Gemini model to write a reply from a prompt, including structured results (an object, a list, or one option from a set of choices) and streamed chat replies.
Actions	Image Generation action	Creates an image from a text prompt with Gemini and returns it to show in the app immediately or upload to storage.
Actions	Embed Text action	Converts a piece of text into a vector embedding with Gemini, for semantic search or grouping similar records by meaning.
Security	Server-side execution	Runs every Gemini call on the WeWeb backend, so prompts and API keys never reach the browser.
Workflows	AI category in backend workflows	Adds Gemini's three actions inside any API Endpoint workflow, alongside your other backend logic.
Usage tracking	AI Requests log	Lists recent Gemini calls with the model used, the cost, and the workflow that ran them.
Usage tracking	AI Budget	Shows how much of the free $1 credit is spent and what's billed this month, with a monthly cap you set in workspace billing.
Environments	Per-environment keys	Uses different Gemini API keys across Editor, Staging, and Production within the same workspace.

Why use WeWeb with Gemini

Gemini provides the intelligence: Google's multimodal models for text, image, and embeddings. WeWeb provides the app, the workflows, and the server-side gateway that call Gemini securely and turn its output into a real product feature.

Because the AI Gateway runs inside WeWeb's native backend, Gemini calls sit next to your existing data. A Text Generation action can read from a WeWeb Table, Supabase, or Xano in the same workflow that calls Gemini, then write the result straight back, no separate glue service required. And because WeWeb apps export as a standard Vue.js codebase, nothing about the Gemini integration locks your frontend to WeWeb: the backend workflows that call Gemini can be self-hosted like the rest of your app.

Perfect for:

AI-powered internal tools that summarize tickets or reports with Gemini
Product description or content generators triggered by a form submission
Customer-facing search that matches by meaning, not exact keywords, using Gemini embeddings
Cover image or mockup generation from a short text brief
Chat assistants that stream Gemini responses into a support widget

Integrating Gemini with WeWeb

Integrating Gemini with WeWeb takes minutes, and Gemini needs no API key to start:

Open Data & API → AI in your WeWeb project. This requires the WeWeb backend; if you don't see the AI tab, set up the backend first.
Add an AI action to a backend workflow and choose Text Generation, Image Generation, or Embed Text.
Leave the provider on Google Gemini (it's the default), or select it explicitly if you've changed it.
Start on WeWeb Cloud, which needs no API key, or go to AI Providers and paste your own Google API key per environment for Gemini usage billed to your own account.
Bind the prompt to workflow inputs or WeWeb variables, then save the response to a table or bind it to the UI.

Learn more about WeWeb's AI Gateway

Best practices

Watch the free $1 credit and your workspace's monthly cap under Data & API → AI → AI Usage; the cap resets on the 1st of the month
Add your own Google API key per environment if you want Gemini usage billed to your Google account instead of the workspace credit; WeWeb does not bill calls that use your own key
Use structured output on the Text Generation action when the next workflow step needs specific fields, not free text
Store generated images in WeWeb Storage or an external bucket if you need to keep them; the Image Generation action does not persist images on its own
Pair the Embed Text action with an external vector database; WeWeb tables cannot store or search embeddings yet

FAQs

1. Does WeWeb integrate with Google Gemini?

Yes. WeWeb's AI Gateway (Data & API → AI) lets backend workflows call Google Gemini for text generation, image generation, and text embeddings, with Gemini set as the default provider.

2. How do I connect Gemini to a WeWeb project?

Open Data & API → AI in a project with the WeWeb backend enabled, add an AI action to a backend workflow, and pick Text Generation, Image Generation, or Embed Text with Google Gemini as the provider. No plugin install and, by default, no API key are required. See WeWeb's AI Gateway docs.

3. Do I need a Google API key to use Gemini in WeWeb?

No. Google Gemini can run through WeWeb Cloud with no API key, using each workspace's free $1 credit and workspace billing after that. OpenAI and Anthropic, by contrast, require your own API key before their actions will run. Details are in the AI Providers docs.

4. What's the difference between WeWeb AI and the Gemini AI Gateway integration?

They're different features. WeWeb AI is the in-editor assistant that helps you build your app (generating UI, workflows, and formulas while you work in the editor). The AI Gateway, covered on this page, is what your published app calls at runtime, for example to summarize a support message or generate a product image with Gemini. WeWeb AI is not required to use Gemini through the AI Gateway, and vice versa.

5. Can I use Gemini to generate images in my WeWeb app?

Yes. The Image Generation action creates an image from a text prompt using Gemini and returns it so you can display it immediately or upload it to storage. See the AI Actions docs.

6. Does WeWeb support streaming Gemini responses?

Yes. The Text Generation action can stream a Gemini reply so the UI shows words as they arrive, useful for chat interfaces where a full response would otherwise feel slow to load.

7. Can I use Gemini for semantic search or embeddings in WeWeb?

Yes, with a caveat. The Embed Text action turns text into a Gemini vector embedding for semantic search or grouping similar records by meaning, but WeWeb tables cannot store or search embeddings yet. You need an external database that supports vectors to store and query the results.

8. Is my Gemini API key or prompt exposed to the browser?

No. Every AI Gateway action, including Gemini calls, runs on the WeWeb server. Prompts and API keys stay server-side and never appear in the browser's network requests.

9. Can I use different Gemini API keys for staging and production?

Yes. Under Data & API → AI → AI Providers, you can set a separate Google API key per environment (Editor, Staging, Production) in the same workspace, or leave any environment on WeWeb Cloud.

10. What happens when the free $1 Gemini credit runs out?

Calls through WeWeb Cloud stop running until you upgrade to a paid plan with auto-charge enabled, raise the monthly cap in workspace billing, or add your own Google API key. The credit and monthly cap only apply to calls made through WeWeb Cloud; calls made with your own API key are never billed by WeWeb. See the AI Usage docs.

11. What are the limits of the native Gemini integration?

The AI Gateway covers three Gemini action types: text generation, image generation, and text embeddings, with common parameters like prompts, structured output, and streaming. It does not expose every Gemini API capability (such as fine-tuning or less common model parameters), and generated embeddings need an external vector database since WeWeb tables don't support vector storage yet. For anything the native actions don't cover, call the Gemini API directly using WeWeb's REST API plugin or a custom backend workflow.

12. Can I use Gemini alongside OpenAI or Anthropic in the same WeWeb project?

Yes. All three providers are available side by side in the same AI Gateway. You can call Gemini for one action and OpenAI or Anthropic for another, each with its own credentials, within the same workspace.

Start building with Gemini today

Add AI to your WeWeb app without managing a separate API integration: Gemini's text, image, and embedding actions are ready to use from your first workflow, no API key required. Get started free and add your first AI action from Data & API → AI.