Bria Product Integrate
Drop a product photo into a scene at exact coordinates — Bria Product Integrate
- scene
- product_image
- IMAGE
- seed
Here's a workflow problem no generic image model solves well: you have a product shot on a clean background and a lifestyle scene, and you want the product sitting in the scene, lit like it belongs there, at a precise position. Bria Product Integrate is the eCommerce answer - you feed it the scene, the product, and an x/y/width/height box, and it composites the product into the scene at exactly those coordinates, with lighting and perspective faked convincingly. It's the node the "shot by" and tailored-generation crowd in this pack is really after.
It's part of the official BRIA AI API nodes pack, hitting BRIA's /v2/image/edit/product/integrate endpoint in the cloud. You need a BRIA API token from platform.bria.ai, no local models, no GPU.
How it works
Both the scene and product_image are base64-encoded and sent to the endpoint, where the product is placed inside the scene at the coordinates you specify, then a generation pass blends it in - adjusting lighting, shadows, and perspective so it doesn't look pasted. The node polls the status URL and returns the composited IMAGE plus the seed used.
The inputs that matter
sceneandproduct_image- required. Scene is where it goes; product is the thing being placed.x_coordinate,y_coordinate,width,height- the placement box. This is the whole control scheme: you're literally telling the model where the product lives and how big it is. Defaults are0, 0, 512, 512- which is the "everything in the top-left corner" trap if you forget to set them.seed- string, default 123456; lock it for repeatable placement.
Outputs: IMAGE (the composite) and seed.
Install
Standard pack install:
cd path_to_comfyui/custom_nodes
git clone https://github.com/Bria-AI/ComfyUI-BRIA-API.git
or ComfyUI Manager → search "BRIA API" → restart, and drop your token in.
The placement box is everything
Unlike the "shot by" nodes which auto-frame the product, this one is manual placement - the coordinates are the feature, and they're also where most people go wrong. A box that doesn't match the product's aspect ratio produces a squashed or stretched composite; a box that's too small leaves the product under-resolved. Measure the product's pixel footprint in the scene and set width/height to match before you spend a single generation credit. And a product with a clean cutout background integrates dramatically better than one with clutter baked in - this node is the last step of a pipeline that starts with Bria RMBG or the eraser, not the first. If you want automatic framing instead, the pack's separate "Shot by Text/Image" nodes are the auto-placement route; Product Integrate is for when you need the pixel position to be yours.
Inputs (8)
| Name | Type | Default | Description |
|---|---|---|---|
| api_token | STRING | BRIA_API_TOKEN | — |
| scene | IMAGE | — | |
| product_image | IMAGE | — | |
| x_coordinate | INT | 00–10000 | — |
| y_coordinate | INT | 00–10000 | — |
| width | INT | 5121–10000 | — |
| height | INT | 5121–10000 | — |
| seedopt | STRING | 123456 | — |
Outputs (2)
| Name | Type | Description |
|---|---|---|
| IMAGE | IMAGE | — |
| seed | INT | — |