LinkedIn multiImage post: 2-20 images, alt text up to 4,086 characters

LinkedIn's multiImage post, created through the Posts API, holds between 2 and 20 images. Each image must be under 36,152,320 pixels and be a JPG, GIF or PNG, and a GIF may have up to 250 frames. Each image can carry altText up to 4,086 characters, though LinkedIn recommends under 120. Generate a consistent set with Sume Images and upload each through the Images API first.

What are the multiImage limits?

These come from LinkedIn's MultiImage Post API page on Microsoft Learn. The images array only supports image URNs of the form urn:li:image:{id}, so each file must be uploaded first.

How do I size the images?

The pixel rule is a total count, not an edge limit. A 3840 by 3840 image has 14,745,600 pixels, comfortably below the cap, so ordinary Sume outputs never come close. Keep every image in the carousel at one aspect_ratio so the slides align; 1:1 and 4:5 are both on Sume's list.

import json, os, urllib.request

body = {
    "model": "bytedance-seed/seedream-4.5",
    "prompt": "slide 1 of 6, bold flat illustration of a rising bar chart, navy and orange, no text",
    "aspect_ratio": "1:1",
    "output_format": "png"
}
req = urllib.request.Request(
    "https://api.sume.com/v1/images",
    data=json.dumps(body).encode(),
    headers={
        "Authorization": "Bearer " + os.environ["SUME_API_KEY"],
        "Content-Type": "application/json",
    },
    method="POST",
)
print(urllib.request.urlopen(req).read().decode())

What goes in the alt text?

Write a short factual description of each slide, close to the recommended 120 characters. If the slide carries words, transcribe them. Sume does not write alt text for you.

What should I verify?

Check the API version header you send against LinkedIn's current version policy. The images guide covers Sume.