Skip to main content
Version: BB AI 1.1

Settings

Configure your Beaver Builder license, AI provider API keys, active model, image generation, Unsplash media integration, layout structure, and chat access in Beaver Builder AI.

Opening the Settings tab​

To open the Beaver Builder AI settings:

  1. Log in to your WordPress admin as an administrator. Beaver Builder AI's settings live in the WordPress Settings menu and are visible to administrators only.
  2. Go to Settings > Beaver Builder AI.
  3. Select the Settings tab.

License​

Enter your Beaver Builder license key to enable automatic plugin updates and to authorize your site for the cloud MCP connection.

License Key​

Your Beaver Builder license key. Paste the key into the field and it is validated automatically.

When connected, a green Connected badge and a Disconnect button appear in place of the input. Click Disconnect to remove the key.

If the Beaver Builder plugin is active and manages its own license, this section shows "Your license is managed by Beaver Builder" with a Manage License link to the Beaver Builder license settings instead of an input field.

AI providers​

An API key for at least one AI provider is required to use the AI chat assistant and generation features. You can connect Anthropic, OpenAI, OpenRouter, and Google Gemini, or point the experimental Custom Endpoint provider at an OpenAI compatible server you run yourself, then choose which one the assistant uses in the Active AI provider section. Usage is billed directly to your account with each provider.

tip

Set a monthly spending limit in each provider's billing dashboard before entering an API key. Beaver Builder AI does not cap usage on your behalf.

danger

If you provide API access for client projects, we recommend creating a separate API key for each client. This makes it easier to manage access, monitor usage, and revoke access when needed.

When a client is no longer using your service, remove or revoke the API key associated with that client. This helps prevent continued access, unexpected usage, or API costs after the client relationship has ended.

Anthropic API Key​

Your Anthropic API key, starting with sk-ant-. The key validates automatically when you paste it or leave the field, by sending a small test request to the provider.

When connected, a green Connected badge and a Disconnect button appear. If validation fails, the error message returned by the provider appears below the field.

Get an API key at console.anthropic.com.

OpenAI API Key​

Your OpenAI API key, starting with sk-. Validation and connection behavior match the Anthropic field.

Get an API key at platform.openai.com.

OpenRouter API Key​

Your OpenRouter API key, starting with sk-or-. OpenRouter provides access to models from many vendors through a single key.

Get an API key at openrouter.ai.

Google Gemini API Key​

Your Google Gemini API key. Validation and connection behavior match the Anthropic field.

A Gemini key does two jobs. It makes Gemini models available as a chat provider, and it powers image generation, which runs through Gemini no matter which chat provider is active. If you only want generated images, connect a Gemini key and leave your active chat provider as it is.

Get an API key at aistudio.google.com.

Custom Endpoint​

An experimental provider for OpenAI compatible servers, such as Ollama and LM Studio, including ones running on your own machine or network. Enter the server's URL, and an API key if the server requires one, then select the provider and one of its models under Active AI provider.

note

The Custom Endpoint provider is experimental. Results depend on the server and model you connect, and smaller local models may not follow the assistant's instructions as reliably as the hosted providers.

Active AI provider​

Choose which connected provider and model the AI assistant uses. Each provider becomes selectable once its API key is connected. Providers without a key are disabled and show "Configure the API key above to enable."

Select a provider, then pick a model from that provider's list:

  • Anthropic offers Claude models, including Claude Fable 5, Claude Opus, Claude Sonnet 5, and Claude Haiku.
  • OpenAI offers GPT models, including the GPT-5.6 family.
  • OpenRouter loads its model catalog from the OpenRouter API and refreshes it daily, giving access to models from many vendors.
  • Google Gemini offers Gemini models.
  • Custom Endpoint lists the models served by the OpenAI compatible server you connected.

Image generation​

Beaver Builder AI can generate on-brand images with Google's Nano Banana image model. Image generation needs a connected Google Gemini API key and runs through Gemini whichever chat provider is active, so it also works while you chat with Claude or another provider.

Once a Gemini key is saved, a Generate with AI option appears on image and background image fields, in the canvas image toolbar, and during page generation. Generated images are saved to your Media Library and come out in the requested shape, such as wide for a hero.

Image fill during page generation

Turn this on to have the assistant generate images for the image spots in a page it builds, instead of filling them with stock photos. It is off by default, since each generated image is billed to your Google account. The model information popover in the chat shows how many images a conversation has generated and what they cost.

See Images for how to use image generation in the editor.

Media providers​

Connecting a media provider gives the AI assistant access to higher-quality stock photography when generating page content. This is optional. The assistant uses Openverse images when no media provider is connected.

danger

If you provide Unsplash API access for client projects, we recommend creating a separate API key for each client. This makes it easier to manage access, monitor usage, and revoke access when needed.

Unsplash API Key

Your Unsplash access key. Paste the key into the field and it validates the connection automatically.

When connected, a green Connected badge and a Disconnect button appear. Click Disconnect to remove the key.

Get a free API key at unsplash.com/developers.

Layout structure​

Choose whether pages built by a connected AI client use rows and columns or boxes. Without this setting the model would choose per section, so a site could end up with a mix of both.

This setting applies to AI clients connected over MCP. The built-in chat generates Custom modules, which do not use either structure. See Native modules for how the two structures differ.

Chat access​

Control where the Beaver Builder AI chat panel appears. Two toggles enable or disable the chat independently for each editor:

  • Enable in Beaver Builder. Shows or hides the AI chat panel inside the Beaver Builder editor.
  • Enable in the Block Editor. Shows or hides the AI chat panel in the WordPress block editor.

Chat is enabled in Beaver Builder by default. Chat in the block editor is enabled by default only when Beaver Builder is not active at the time the plugin is activated.

When enabled, the chat is available to users with the unfiltered_html capability, which typically includes administrators and editors on single-site installs.