Imagen on Vertex AI | AI Image Generator

To see an example of getting started with creating high quality visual assets with Imagen and Gemini, run the "Create High Quality Visual Assets with Imagen and Gemini" Jupyter notebook in one of the following environments:

Open in Colab | Open in Colab Enterprise | Open in Vertex AI Workbench user-managed notebooks | View on GitHub

API reference overview: To view an overview of the API options for image generation and editing, see the imagegeneration model API reference.

Imagen on Vertex AI brings Google's state of the art image generative AI capabilities to application developers. With Imagen on Vertex AI, application developers can build next-generation AI products that transform their user's imagination into high quality visual assets using AI generation, in seconds.

With Imagen, you can do the following:

• Generate novel images using only a text prompt (text-to-image AI generation).
• Edit an entire uploaded or generated image with a text prompt.
• Edit only parts of an uploaded or generated image using a mask area you define.
• Upscale existing, generated, or edited images.
• Fine-tune a model with a specific subject (for example, a specific handbag or shoe) for image generation.

^{Image generated with Imagen on Vertex AI from the prompt: portrait of a french bulldog at the beach, 85mm f/2.8 (Imagen 3 model version).}

Product usage

To view usage standards and content restrictions associated with Imagen on Vertex AI, see the usage guidelines.

Person and face generation

Imagen 3 generates the most realistic and highest quality images from natural language text prompts, including images of people of all ages. These person generation features, including the generation of adult and child images, may require your use case to be reviewed and approved.

Imagen 3 may provide you an error that indicates that your Google Cloud project needs to be approved for either adult generation or child generation, depending on the person or face generation parameter that you choose and the context of your text prompt.

If you require approval, fill out the request form, and a Google representative will follow up on your request.

Imagen on Vertex AI features and launch stage

Imagen on Vertex AI offers several image generative AI features. These features are available at different launch stages. The features and stages are as follows:

Feature	General Availability	General Availability (requires approval^*)	Preview
Image generation^* with Imagen 3 and Imagen 3 Fast	✔
Digital watermarking and verification (image generation)	✔
User-configurable safety settings (image generation)	✔
Imagen 3 Customization^* (few-shot learning): Subject customization, Style customization, Controlled customization, and Instruct customization.		✔
Imagen 3 image editing (mask-based: inpainting insert or remove, outpainting, product image editing)^*		✔
Imagen image editing (mask-based and mask-free)^*		✔
Edit using Imagen 2 Personalization^*			✔
Edit using Imagen 2 Controlled Customization^*			✔
Subject model fine-tuning (standard)^*		✔
Subject model fine-tuning (instant)^*			✔
Custom style model training^*		✔
Visual captioning	✔
Visual Question Answering (VQA)	✔
Text-to-live images^†			✔
Video description^*			✔

^* To request access to use this Imagen feature, fill out the Imagen on Vertex AI access request form.
^† Contact your account representative about gaining access to this feature.

List of articles with samples

Use the following articles to understand Imagen on Vertex AI use cases.

For Jupyter notebook tutorials using Imagen, see Generative AI on Vertex AI notebook tutorials.

Features	Description	Links
Image generation (Imagen 3)	A developer's guide to getting started with Imagen 3 on Vertex AI. Read about new Imagen 3 model features. Learn more about Imagen's image generation feature.	Article link
Image generation (Imagen 3)	Do It Yourself Imagen 3 - Practical Demo with Vertex AI. Run a Colab that uses new Imagen 3 and Imagen 3 Fast model features. Learn more about Imagen's image generation feature.	Article link
Image generation (Imagen 2) Image editing (Imagen)	Image generation with Imagen and LangChain4j (Java). In this article, we will explore how you can generate and edit images with Imagen in LangChain4j. Learn more about Imagen's image generation and image editing features.	Article link
Image editing (Imagen)	Background change with Imagen on Vertex AI: A step-by-step guide. This article shows how to do mask-based editing using the Vertex AI SDK for Python. Changing the background involves the following steps: Remove the existing background. Create a mask and inverted mask image. Encode an image to a string. Create a request payload. Learn more about Imagen's image editing feature.	Article link GitHub
Image generation (Imagen) Text generation (PaLM 2 for Text)	Google Imagen (through Google Cloud Vertex AI Studio) as fashion design assistant. In this article, we will explore how generative AI can assist fashion designers in generating new ideas and designs using Google's suite of generative models for text and image generation. This article shows you how to use PaLM 2 for Text's text-bison model in the Google Cloud console and with the Vertex AI SDK for Python. It then shows you how to use the generated prompts to serve as input for image generation using Imagen's imagegeneration model. Learn more about Imagen's image generation feature and PaLM 2 for Text's text generation feature.	Article link

Restricted GA and Preview features

Imagen on Vertex AI offers limited access GA features and Preview features available under our Trusted Tester Program.

To request access to use these restricted Imagen features, fill out the Imagen on Vertex AI access request form. If approved, you will be provided with instructions on how to get started.

What's next

Use the following links to view the feature documentation.

Prompt guide

See how to write effective prompts to generate images.

How-to: Generate images

Learn how to generate images with Imagen on Vertex AI.

API reference: Generate images

Learn about optional and required fields when sending an Imagen image generation request.

^{Image credit: All images generated using Imagen on Vertex AI.}