Spin-off for an AI series: reuse a character sheet in a new show
A spin-off keeps one character and changes the world. Generate a new sheet with up to 16 references, then reuse the same stills across the new show's episodes.

To spin a character off into a new show, take the best stills of that character from the parent series, use them as references to make a new character sheet in the new setting, and then send that sheet as the reference in every episode of the new show. Sume's ChatGPT Image 2.5 models accept up to 16 image references, so the old look and the new costume or location can share one request.
Why a sheet and not the old clips
A spin-off changes the setting but not the face, and a model drifts less when it is shown the face in the same few poses every time. A character sheet is those few poses on one image. The parent series already gave you the raw material: pull stills from your own finished episodes with the video frames endpoint, which returns durable media.sume.com images at the source size.
Build the new sheet
Send the strongest stills as input_references to openai/gpt-image-2.5 together with a prompt that names the new setting, outfit and lighting. Ask for a front view, a three-quarter view and a profile in the same image or in three images, and keep the style sentence identical across the three. Ideogram 4.5 is also on Sume and takes up to 5 images in total when you give it references, which is a tighter fit when you only have a few stills.
| Surface | What it accepts |
|---|---|
ChatGPT Image 2.5 (openai/gpt-image-2.5) | Up to 16 image references |
Ideogram 4.5 (ideogram/ideogram-v4.5) | Edits the first image and uses up to 4 more, 5 in total |
Format run attachments | Up to 30 images; media URLs in input share the same budget |
| Reference URLs | Public HTTPS only; localhost, private-network and non-HTTPS URLs are rejected |
Use it in the new show
Send the sheet as a reference in each video request, and also send a scene still as the first_frame when you need an exact opening. The two fields work differently: input_references guides the look, and frame_images pins a frame. If you send both, Sume treats the request as image-to-video, so decide which job each episode needs. Keep the sheet URL in your season config so every episode's request is built from the same file.
Keep the two shows apart
- Give the spin-off its own season config, its own Idempotency-Key prefix and its own folder of references, so a change to the parent's look never reaches it by accident.
- Change one thing the audience can see in the first five seconds of episode 1: a new location or a new colour. A spin-off that looks like a rerun is harder to explain.
- Use only characters, faces and voices you have the right to use. A spin-off of someone else's character is their decision.
Sources
Related posts
More in Use cases
- AI spokesperson release notes, 90 seconds: two jobs, cost by tier
A 90-second spokesperson video is two Sume avatar jobs because one job tops out at 60 seconds. Cost at standard, plus and max, and how to split the script.
- AI UGC creator package: cost basis for 3, 5 or 10 videos
Selling AI UGC videos? One $0.95 avatar and 20-second product clips cost $12.59 for three, $20.35 for five and $39.75 for ten at standard quality.
- How do I make a spoken wake-up alarm with an AI voice and music?
A wake-up alarm is a short voice greeting joined to a music file: 2 cents of TTS and a 1-cent concat per day, with one $0.125 music track reused all week.
- Airbnb house-rules video: an avatar in front of your living room
A short-term rental host can make a 9:16 house-rules video by placing a Sume avatar in front of a photo of the unit, using the talking-video photo scene input.
Written by Sume