Skip to main content
The brand-aware generation workflow below uses a ready brand and applies its current context automatically. Your request describes the subject, composition, and medium needed for the image.

​
Prepare the brand and prompt

Use an existing brand whose status is ready, or create a brand and wait for it. Pass its Brand ID as brandSessionId in the API or brand_session_id in MCP. Describe what should appear in the image:
  • the subject and important objects;
  • the composition or framing;
  • the medium, such as a photograph, illustration, 3D render, mockup, or poster; and
  • any text or product detail that must be present.
Bloom already applies the brand’s visual guidance. Avoid generic aesthetic filler such as “professional,” “stunning,” or “premium.”

​
Start generation

Send POST /images/generations:
Both interfaces return image IDs while generation continues. Keep those IDs to retrieve the results. The generation reference and live MCP schema define the exact request and response fields.

​
Retrieve the result

Retrieve one image with GET /images/{id}?wait=true:
For several image IDs, use GET /images?ids=...&wait=true to collect them together.
A wait can end before generation finishes. Repeat the same read until every image is completed or failed. Use the returned image URL only after completion; if generation fails, read the failure information before retrying.
Landscape rabbit r1 campaign concept with a hand holding the orange device and the headline agents on the go

An independent rabbit campaign example created with Bloom and adapted to 16:9. This is not an official rabbit campaign.

Follow the campaign image workflow to see how an agent used brand references, explored three directions, and adapted the selected image to multiple sizes.

​
Use reference images

References can ground the generation in a specific product, person, composition, or visual example. With the API, pass up to 10 uploaded or generated image IDs as referenceImageIds. Use the image upload or Brand Library search endpoints to obtain those IDs. With MCP, use search_user_images to search the Brand Library or upload_image to add an image by URL. Pass up to 10 selected IDs as reference_image_ids to generate_image. For each generation, include only the references relevant to that image and describe how each should be used. Refer to a reference by its content, not its position in the input list.

​
Use reference ads

Through MCP, ask your agent to search the reference-ad library and show options before adapting one to your brand:
The selected ad determines the composition and aspect ratio. See the MCP tool overview for the tools involved.

​
Choose a model

The example uses pro for final work. Use the generation reference or your live MCP schema to choose a supported model, size, ratio, and variant count. See Credits and limits for costs. Model Gateway generation is a separate path and does not apply brand context automatically. If your agent uses generate_image_with_model, it must read relevant guidance and files with get_brand and view_files, then include that context in the model’s supported inputs. See Model discovery and generation for availability; the live MCP schema defines the exact inputs. Once an image is complete, continue with Edit and adapt images.