AI car color preview: try a wrap or paint color on your car photo

Mask the body, keep the glass and wheels

A wrap shop or an owner deciding between colors wants to see the real car in the new color. The car photo is the reference; a mask over the body panels tells the edit where to work, and the prompt names the color and finish.

Use mask_url with ChatGPT Image 2.5, since Sume documents the field on that model only. Without a mask the prompt alone has to hold back the edit, and wheels, glass and badges are the first things to drift.

Sume Image API docs list ChatGPT Image 2.5 as openai/gpt-image-2.5 (Flare) and openai/gpt-image-2.5-sunburst. OpenAI's guide says to choose Sunburst where editing precision matters most and Flare for fast everyday generation, so these edits use Sunburst.

Name the finish and the parts to keep

Say gloss, satin or matte, because finish changes how the body reads. List what must stay exactly as photographed: windows, wheels, tires, lights, badges, number plate and the background. If the plate or badges carry text, check them at full size.

import os
import requests

REFS = [
    "https://example.com/car.jpg",
]
resp = requests.post(
    "https://api.sume.com/v1/images",
    headers={"Authorization": f"Bearer {os.environ['SUME_API_KEY']}"},
    json={
        "model": "openai/gpt-image-2.5-sunburst",
        "prompt": "Recolor the body panels satin dark green. Keep windows, "
                  "wheels, tires, lights, badges, number plate and the "
                  "background unchanged.",
        "aspect_ratio": "auto",
        "mask_url": "https://example.com/car-body-mask.png",
        "input_references": [
            {"type": "image_url", "image_url": {"url": u}} for u in REFS
        ],
    },
    timeout=60,
)
print(resp.status_code)
print(resp.json())

Compare finishes in separate calls

Run one call per color and finish with the same photo and mask. The mask rules from OpenAI's guide apply: alpha channel, same size and format as the photo, under 50 MB.

Do not treat it as a wrap proof

A wrap installer works from a film sample on the actual panel. Use the render to narrow the list to one or two colors, then ask for a physical sample in daylight.

Quality and the first try

On ChatGPT Image 2.5 the quality field takes auto, low, medium, high, xhigh or max, and leaving it out means high. For a first pass at a layout idea, a lower tier is a reasonable way to look at composition before you pay for a final render.

Keep the source photo, the prompt and the response together for each option. That makes it easy to rerun the one you pick at a higher quality tier.

Sync, jobs and the bill

Treat the response code as the switch. 200 means the image body is in the response. 202 means a job was created because the 30-second wait ran out, and the image is read later from GET /v1/jobs/{id}/result.

A completed image is billed in full and a failed or cancelled one is not. The charge shown in usage.cost is provider list price times 1.25, so you can log it per edit and sum a batch from those numbers.