Shot by Image - Automatic
Get seven compositions, pick the winner
- image
- ref_image
- output_image_1
- output_image_2
- output_image_3
- output_image_4
- output_image_5
- output_image_6
- output_image_7
ShotByImageAutomatic (shown as Shot by Image - Automatic) is the "let Bria art-direct" option in the product-shot family. Same engine as ShotByImageOriginal - product photo in, scene built from a reference image - except instead of preserving your product's composition, the model decides where the product should sit, how big it should be, and what the scene should look like. And it doesn't hand you one guess: it returns up to seven different compositions on seven output sockets.
If the "Original" variant is for when your product photo is already framed right, this is for when you don't know what looks good yet. Fire it at a product on a white background, get seven candidate catalog shots, pick the one that sells.
How it works
Inputs: image (product) and ref_image (scene reference), plus shot_size (default 1000, 1000) - the canvas dimensions for the output shots, given as width, height in pixels. The node calls Bria's /v1/product/lifestyle_shot_by_image endpoint with the automatic placement type, which tells the model to compose freely. The shared controls apply: enhance_ref_image (default true), ref_image_influence (0–1, default 1.0), force_rmbg, content_moderation.
The output is the whole point: output_image_1 through output_image_7, each an IMAGE. The API only returns as many as it generates (up to 7); the node pads the rest with None, so unused sockets just sit empty - that's normal, not a bug. Pick the best one and wire it to Preview/Save.
Installing it
Part of the BRIA AI API nodes pack (Bria-AI/ComfyUI-BRIA-API). ComfyUI Manager → search BRIA API → install, or:
cd ComfyUI/custom_nodes
git clone https://github.com/Bria-AI/ComfyUI-BRIA-API
Restart ComfyUI; it's under API Nodes. No model downloads, no GPU - cloud call needing a BRIA API token from platform.bria.ai. Each run is a single billed generation, and the seven outputs come from that one call, so you're not paying seven times for the batch. That's the value pitch in one sentence.
Practical notes
Seven candidates sounds like a gift, but watch two things. First, the returned set can be inconsistent in quality - that's the nature of "automatic"; you're paying for variety, so budget time to actually review all seven. Second, shot_size is your only layout constraint here, so a weird aspect ratio gets you weird compositions. If you want more predictable results, the Manual Placement and Custom Coordinates variants pin the product down; if you want the same free-composition but with an aspect ratio instead of pixel dimensions, use ShotByImageAutomaticAspectRatio.
Inputs (8)
| Name | Type | Default | Description |
|---|---|---|---|
| api_key | STRING | BRIA_API_TOKEN | — |
| image | IMAGE | — | |
| ref_image | IMAGE | — | |
| shot_size | STRING | 1000, 1000 | — |
| force_rmbgopt | BOOLEAN | false | — |
| content_moderationopt | BOOLEAN | false | — |
| enhance_ref_imageopt | BOOLEAN | true | — |
| ref_image_influenceopt | FLOAT | 1.000–1 | — |
Outputs (7)
| Name | Type | Description |
|---|---|---|
| output_image_1 | IMAGE | — |
| output_image_2 | IMAGE | — |
| output_image_3 | IMAGE | — |
| output_image_4 | IMAGE | — |
| output_image_5 | IMAGE | — |
| output_image_6 | IMAGE | — |
| output_image_7 | IMAGE | — |