Digital signage greeter: a 9:16 avatar clip loop, priced on Sume

For a digital signage greeter, render a short avatar clip per message with aspect_ratio set to the screen's orientation (9:16 is the default, and 1:1, 3:4, 4:3 and 16:9 are also accepted), then loop the MP4s in your signage player. Five 20-second clips cost $18.40 on Standard, because Sume prices by seconds of video and the aspect ratio does not change the rate.

This is a rendered rotation, not a live agent: a person walking up cannot ask it a question. Facts are from Sume's avatar video guide and the public catalog, read 2026-10-04.

Pick the aspect ratio per screen

A portrait totem and a landscape wall screen need different frames, so submit one job per orientation. The script and avatar stay the same; only aspect_ratio changes. Resolution is currently 720p for every ratio, so check that a large wall screen is not scaling a clip well past its size before you commit.

One rotation, priced

Plan the loop as short clips of one idea each: a welcome, today's hours, a wayfinding tip, a promotion, a closing. Each is its own job, so changing one message later means re-rendering one clip, not the whole loop.

Submit a clip

The call below renders one portrait greeting. Replace product_host with a ready avatar handle in your workspace.

curl -X POST https://api.sume.com/v1/avatar-1.0/talking-video \
  -H "Authorization: Bearer $SUME_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: lobby-greeting-portrait-001" \
  -d '{"avatar_handle": "product_host", "aspect_ratio": "9:16", "quality": "standard", "script": "Welcome. The front desk is straight ahead, and today's visitor badges are on the left."}'

Keep the loop honest

Signage runs unattended for months, so a clip with a stale hour or a past promotion is a quiet failure. Store each clip's media.sume.com URL with a date you review it by, and re-render rather than editing the file.

Say in the clip, in the caption or in the message around it that the presenter is AI-generated. The API does not add that label for you, so put it in the script or the page the video sits on.