Docs

Torch Object Storage

User Guide

TOS-Doubao AIGC

TOS now works with Doubao's large models on Volcano Ark, allowing you to work across different platforms. TOS's AIGC (AI-Generated Content) APIs combine Douban's AIGC capabilities with TOS's image processing features to meet diverse business needs. This integration lets you create AI images while also doing things like resizing and converting formats. This topic explains how to use AIGC APIs to process images. It includes the capabilities, billing details, and operational steps of this feature.

Capabilities

Doubao's text-to-image model creates highly detailed and accurate images based on text descriptions. The image-to-image model can produce images in various creative styles. These models can be used to customize avatars and make photo albums. For further details, see Text-to-image and image-to-image capabilities. Doubao's large models have some current limits on image formats, file sizes, and other aspects. You may need to preprocess images to meet the model's requirements before entering the prompts. Also, changes may be needed to the model's results to match what the business needs.

TOS offers strong support for Doubao's large models with its efficient data storage and flexible image processing features, such as image resizing, format conversion, and brightness adjustment. You can combine TOS's image processing features with Doubao's text-to-image and image-to-image models through AIGC APIs to meet diverse business demands.

Use cases

TOS offers AIGC capabilities based on Doubao's large models. Here are some examples of how this could be used:

Social media content: Make interesting visual content for social media to get people to interact and pay attention.

Advertising creativity: Create advertising images quickly and easily. These images will help marketing teams try out different ad designs.

Teaching assistance: Create visual aids for complicated educational content (like historical events and scientific concepts) to help students understand and remember the material.

Artistic creation: Artists can use text-to-image technology to try out new styles and ways to express themselves, creating unique works of art.

Film and game design: Before they start designing a film or video game, creative teams can use text-to-image technology to quickly create scene sketches or character designs. This helps them develop more concrete visual ideas.

Billing notes

Using AIGC APIs incurs costs for smart drawing API calls and image processing. TOS does not charge for smart drawing API calls. This cost is managed by Doubao. Additionally, Doubao gives each account a certain amount of free smart drawing quota. TOS may charge you for image processing when you use AIGC APIs. You will get 10 TiB for free each month. If you go over this limit, TOS will charge you based on the size of the original images.

How it works

TOS supports three different ways to use the image processing features and Doubao's large models:

Image preprocessing: First, use TOS's image processing features (e.g., resizing and formatting) to prepare the original image. Then, use Doubao's large model to generate the new image you want.

Image postprocessing: First, use Doubao's large model to create an image, then improve it with TOS's image processing features (e.g., format, brightness, contrast adjustments).

Image preprocessing & postprocessing: First, use TOS's image processing features (e.g., resizing and formatting) to prepare the original image. Then, use Doubao's large model to generate the new image you want. Finally, use TOS's image processing features (e.g., format, brightness, contrast adjustments) again to improve the result.

Steps

Activate and authorize Doubao's smart drawing models.

In the left navigation pane, click AIGC.

On the AIGC page, under Model activation, click Activate to enable the smart drawing model.

Under Model authorization, click Authorize to allow TOS to use the smart drawing model.

Under Object Storage Authorization, click Authorize to allow access to files in your TOS bucket.

Call AIGC APIs to combine TOS's image processing features with Doubao's large models.

The table below shows the documentation, usage, and large models available for each AIGC API. Use the instructions to make your API calls.

tip

Currently, you can only combine TOS's image processing features with Doubao's large model via API calls. You cannot do this through the TOS console.

AIGC API documentation	Usage	Available Doubao large models
TextToImage	Image postprocessing	Doubao-2.1-drawing-t2i Doubao-2.0-drawing-pro-t2i Doubao-2.0-drawing-t2i Doubao-XL-drawing-pro-t2i (text-to-image) Doubao-1.4-drawing-t2i Doubao-1.3.X-drawing-lite-t2i
ImageToImage	Image preprocessing	Doubao-XL-drawing-pro-t2i (image-to-image)
			Image postprocessing
			Image preprocessing & postprocessing

Last updated: 2025.09.25 17:12:49