Generate images with AI

Odoo AI can generate images directly from natural language prompts. AI-generated images can be used throughout the database for product visuals, website content, marketing campaigns, Knowledge articles, and other business workflows.

Image generation is available in supported image fields and editors, allowing visuals to be created and refined directly within Odoo.

Generate an image

To generate an image, navigate to the AI app and click on the Odoo Image Generation Agent to open a conversation. Enter a prompt describing the desired image, then click enter.

The prompt can describe the subject, environment, lighting, perspective, style, or intended usage scenario. More detailed prompts generally produce more accurate and consistent results.

After processing is complete, the generated image appears directly in the conversation window. If the generated image does not match the intended result, the prompt can be refined and the image regenerated.

An example conversation with the agent.

Important

If text, branding, or logos are required in the generated image, these elements must be explicitly included in the prompt. Otherwise, generated images exclude visible text by default, according to the rules defined in both the image generation agent and the Generate Images topic.

Generate an image on a webpage

To generate an image to add directly to a webpage, navigate to the Website app and click Edit. Either click on an existing block or add a new one.

In the webpage editor, under the Image drop-down menu, click Replace next to the Media field to open the Select a media pop-up. On the pop-up, click the AI icon.

The Select a media pop-up window.

On the resulting conversation window, either click one of the default prompts, or enter a custom prompt.

Write effective prompts

Prompts describe the subject and context of the generated image. More detailed prompts generally produce more accurate and consistent results.

Effective prompts often include:

  • The subject of the image.

  • The environment or setting.

  • The perspective or composition.

  • The intended visual style.

For example, instead of using a short prompt such as, Office Chair a more descriptive prompt may produce stronger results:

Modern ergonomic office chair in a collaborative workspace with natural daylight and a side-angle product photography composition

When generating variations from an existing image, prompts work best when describing how the subject should be presented or used, rather than redefining the subject itself.

How AI image generation works

AI image generation in Odoo is powered by a dedicated AI agent called the Odoo Image Generation Agent. This agent interprets prompts, generates images, and applies the rules that guide image quality, consistency, and behavior.

The image generation agent works together with a topic called Generate Images. The agent prompt defines the high-level behavior of the image generation workflow, while the topic instructions provide more detailed guidance regarding how images should be generated and refined.

The agent prompt primarily controls how text and branding are handled in generated images. By default, generated images do not contain visible text, labels, logos, signage, or watermarks unless these are explicitly requested in the prompt. This helps reduce unintended visual artifacts and improves overall image consistency.

The Generate Images topic contains the detailed image-generation instructions used during the creation process. These instructions guide how the AI should preserve the identity of a subject, how new environments should be generated, and how each image variation should remain visually distinct from previous generations.

Together, the agent and topic ensure that generated images remain both visually coherent and contextually relevant while still allowing for creative variation.