Ideogram 4.5 on Sume does not accept output_format

Most Sume image ids accept output_format as png, jpeg or webp. Ideogram 4.5 does not. The Sume API notes state that its provider selects the output format and the API does not accept the field, and its catalog descriptor lists no formats. Recraft V4 is the other special case: its only listed format is webp.

Ideogram's own docs list JPEG, PNG and WebP for uploaded reference images (25 MB each), which is about inputs, not what you get back.

Formats by id

Read the type that came back

Each item in data[] has a media_type, so branch on it instead of assuming PNG.

import io, os, requests
from PIL import Image
headers = {"Authorization": "Bearer " + os.environ["SUME_API_KEY"]}
r = requests.post("https://api.sume.com/v1/images", headers=headers, timeout=60,
                  json={"model": "ideogram/ideogram-v4.5", "prompt": "Menu board"})
item = r.json()["data"][0]
print(item["media_type"])
img = Image.open(io.BytesIO(requests.get(item["url"], timeout=30).content))
img.convert("RGB").save("out.jpg", quality=90)

Transparent output

Output format also matters for transparency. On Sume, background: transparent is accepted only on ChatGPT Image 2.5, and OpenAI's guide says to pair it with png or webp.

How this was checked

Vendor facts come from the pages listed in the sources, read on 2026-10-05. Sume facts come from the Image API docs and the catalog code on main on the same date. Catalogs and limits change, so read the descriptors from GET /v1/images/models before you pin a number in production code.