10 FAQ answers as 15 second avatar clips: $27.60 on Sume standard
A clip bank of ten FAQ answers, 15 seconds each, from one avatar handle: $27.60 on standard, $36.75 on plus, $82.50 on max, plus $0.95 once for the avatar.
Ten FAQ answers as 15 second avatar clips cost $27.60 on Sume standard quality, $36.75 on plus and $82.50 on max, after a one-time $0.95 to create the avatar. That is 150 seconds of video at $0.184, $0.245 or $0.55 per second. A clip bank like this is the practical stand-in for a live conversational avatar while models such as Tavus Griffin-Lite are limited to select trusted testers.
| Quality | Arithmetic | Clips total | With $0.95 avatar |
|---|---|---|---|
| standard | 150 x 0.184 | $27.60 | $28.55 |
| plus (default) | 150 x 0.245 | $36.75 | $37.70 |
| max | 150 x 0.55 | $82.50 | $83.45 |
Why 15 seconds
A 15 second answer is about 42 words at the 2.8 words per second Sume uses to plan clips. That is long enough for a direct answer and one next step, and short enough that people read the question, press play and leave with the answer. The job window is 4 to 60 seconds, so 15 is well inside it.
Write the answers as a list first. The first sentence should repeat the question in the viewer's words, because a clip is played out of context from a help page.
Building it
Create the avatar once from a prompt, props or a photo with POST /v1/avatar-1.0/generate and a handle such as faq_host. Wait for the job to complete. Then send ten talking-video requests with the same avatar_handle, each with its own Idempotency-Key, so a retry never creates a second paid job.
- Use the same
aspect_ratiofor all ten, 16:9 if they embed in a help page. - Set
qualityexplicitly on every request, because the default is plus. - Store the job id, the handle and the approved script text per answer, so you can find which clip to redo when the answer changes.
What this does not do
A clip bank answers the questions you wrote. It does not listen, interrupt or adapt to a follow-up, and the avatar speaks English because the talking-video prompt allows no other language. When a question is not in the bank, send the visitor to text support rather than showing the nearest clip.
Changing one answer costs one clip. Re-rendering a 15 second answer on standard is 15 x 0.184 = $2.76.
Keeping the bank current
Answers go stale. Keep one row per clip: question, approved script, quality tier, job id, result URL and the date you last checked the answer. When a policy changes, find the rows that mention it and redo only those. If three of the ten change a quarter, the refresh cost on standard is 3 x 2.76 = $8.28.
Embed each clip next to its written answer, not instead of it. Search engines and screen readers need the text, and some visitors would rather read.
Sources
Related posts
More in Use cases
- Test an avatar's look with a 4 second clip before a 60 second run
A 4 second Sume clip costs $0.736 on standard; a 60 second one $11.04. Use the cheap run to check face, voice and framing first. Numbers and checklist.
- Test camera control on Sume: five 5-second prompts for $3.13 a row
Vidu Q4 Preview claims stable camera movement. Five 5-second camera prompts cost $3.125 on Wan 3.0 or Omni at 720p and $2.50 on MiniMax H3 Max at 768p.
- Test five presenters on one 9-second ad: Recast $16.90 at 768p
Five presenter swaps of one 9-second ad cost $16.90 at 768p or $25.35 at 1080p with H3 Max Recast on Sume. Per-job rounding and a testing plan inside.
- 3 GPT low drafts, then a Nano Banana 2.1 2K final: 22 cents on Sume
Three GPT Image 2.5 low drafts plus one Nano Banana 2.1 2K final cost 22 cents on Sume ($0.07425 + $0.15). A cheap way to pick a composition first.
Written by Sume