Appcues flow images: 5 MB cap, 650 px modals, Sume hero art

Appcues says JPG, PNG, SVG and GIF are supported with a 5 MB maximum, and under 1 MB is recommended. Modals want at least 650 px width. Ask Sume for a JPEG at a size near that and check the byte count before you upload to Appcues.

Sizes are from Add Text, Images and GIFs to Experiences, read 2026-10-03.

What sizes does Appcues recommend?

The hero image is cropped to fit, so keep the subject away from the top and bottom edges.

How do I generate one with Sume?

Image Router takes aspect_ratio or, on GPT image models, a custom image_size: both edges multiples of 16, at most 3840, ratio at most 3:1, and 655,360 to 8,294,400 pixels. 1300 by 650 fails the multiples-of-16 rule on 1300, while 1296 by 656 passes. Models accept only ratios in their own supported_parameters, so check GET /v1/images/models first.

curl -X POST https://api.sume.com/v1/images \
  -H "Authorization: Bearer $SUME_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: appcues-hero-001" \
  -d '{
    "model": "sume/auto",
    "prompt": "Soft abstract dashboard shapes, calm blue, plain edges, no text",
    "aspect_ratio": "2:1",
    "output_format": "jpeg"
  }'

How do I stay under the 5 MB cap?

Sume does not promise a file size, so read the finished artifact's size after the job and compare with Appcues' 5 MB limit and its under-1 MB recommendation. If it is too big, request again at a smaller size. Keep text out of the generated pixels; put words in Appcues' own text component, where they are exact.

What does Sume not do?

No Appcues connector appears in Sume's docs, so you upload the file. Sume also does not tell you how a modal will crop an image; preview it in the Appcues builder.