New-hire welcome video: 30-second avatar with auto captions
A 30-second welcome video from an Avatar 1.0 talking video plus speech captions costs $5.72 on Sume. Chain, limits and English-only caveat, read 2026-10-10.
A 30-second welcome video from an Avatar 1.0 talking video at standard quality costs $5.52, and burning captions from its speech adds $0.20, for $5.72 once the avatar exists. Avatar creation is a one-time $0.95. Scripts must plan to 4 to 60 seconds and are English-only.
HR and team leads send the same welcome many times with small changes. The reusable avatar makes each version a script edit. Captions matter because new hires often watch muted at a desk. Prices come from the Sume pricing code (read 2026-10-10).
What is the endpoint chain?
Two writes after a one-time setup.
- Once:
POST /v1/avatar-1.0/generatewith a photo to create the presenter and save theavatar_handle. POST /v1/avatar-1.0/talking-videowith the handle, a 30-second script (roughly 75 to 80 words),quality: standard,aspect_ratio: 16:9.POST /v1/video-captionson the result with no cue fields, so Sume transcribes the speech, and a style you pick for the whole team's videos.- Optional:
POST /v1/video-trimat $0.02 to cut a long take down.
What does it cost by quality and length?
Avatar seconds are the main cost, billed per second by quality.
| Length | Standard $0.184/s | Plus $0.245/s | Max $0.55/s |
|---|---|---|---|
| 20 s | $3.88 | $5.10 | $11.20 |
| 30 s | $5.72 | $7.55 | $16.70 |
| 45 s | $8.48 | $11.22 | $24.95 |
What are the limits?
Avatar 1.0 is English-only, and a script that plans under 4 seconds or over 60 is refused. Voice cloning is not an API feature, so the voice is a catalogue voice, not the manager's. If your hires speak another language, add translated captions as cues, not a translated avatar; the face will still speak English.
Do not generate a face of a real person without their permission. An avatar made from an employee's photo needs that employee's consent, and the video must not suggest they said something they did not. See Generate avatar video and Create new avatar for input rules.
Keep policy details out of the video: they change, and the cost to re-render is the whole table again.
When is a recorded video better?
If the manager can record 30 seconds on a phone, that is free and more personal. Use the avatar for repeatable parts, like compliance reminders or a role-specific tour. Related price planning is in AI avatar for online course videos and the training explainer plan-first note.
Pre-publish checklist
Read the script aloud once. Avatar time is the only expensive part, so catch awkward lines before you generate, not after.
Plan updates: when a name or date changes, the whole 30 seconds is regenerated at $5.52, so keep volatile details out of the script.
- Keep the script under 80 words.
- Use plain English the avatar can pronounce.
- Add team names as caption cues rather than speech.
Sources
Related posts
More in Use cases
- Newsletter to video: sentence-timed TTS drives four slots, 74 cents
Turn a 900-character newsletter intro into a video on Sume: sentence segments set each slot start, four stills, one render, one caption job (read 2026-10-10).
- One 16:9 hero video cropped to 9:16 and 1:1 for $1.04
Cut a 16:9 Wan 3.0 hero video into 9:16 and 1:1 versions with Sume Video filter crop fractions at $0.02 each. Chain, math and limits, read 2026-10-10.
- One 30-second Wan take cut into three variants with Video trim
Generate one 30-second Wan 3.0 take at 480p, then cut three variants with Video trim at $0.02 each. Chain, cost and limits, read 2026-10-10.
- One packshot to twelve lifestyle scenes: input_references recipe
Turn a single product packshot into lifestyle scenes with the Sume Image API: which models take it as a reference, the request body, and cost per twelve scenes.
Written by Sume