School closing announcement as a 30-second avatar video with captions

A school closing notice fits a 30-second avatar video in the default 9:16 frame, which suits phone screens, with captions burned in for parents who watch muted. It costs $5.72 on Standard, $7.55 on Plus and $16.70 on Max including the fixed $0.20 caption add-on, and Standard is the fastest execution path.

Sume does not promise a delivery time for a job, and the docs describe queued and processing states, so do not plan to render a same-morning notice after the decision is announced. Facts are from the avatar video guide and Jobs and results, read 2026-10-04.

When this is a good idea

Predictable communications are a good fit: a weekly update, an enrollment reminder, an event recap. A real-time emergency notice is not, because an unattended job queue is the wrong place for a decision that must reach people in minutes. Use your existing alert channel for that, and use video as a follow-up.

Tiers and cost

A script with the facts parents need

Open with the decision, then the date, then what happens next. Keep to what a family needs to act on. Parents read captions, so spell the school name the way you want it shown.

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: school-closing-2026-01-14" \
  -d '{"avatar_handle": "school_office", "aspect_ratio": "9:16", "quality": "standard", "script": "Good evening. Oak Grove school will be closed tomorrow, Wednesday, because of the snow. Classes will resume on Thursday on the usual schedule. Afternoon activities are cancelled. Please check the school app for any update tonight.", "captions": {"enabled": true, "style": "slam", "language": "en"}}'

Consent and honesty

If the avatar looks like a staff member, get that person's written permission before you create it from their photo, and tell families it is a generated presenter. Say in the clip, the caption or the page around it that the presenter is AI-generated; the API does not add that label for you.