AI presenter video for job postings: 25 roles, 20 seconds each
25 twenty-second hiring clips from one recruiter avatar cost $122.50 on Plus or $92.00 on Standard, plus $0.95 once, at Sume's listed rates.

Twenty-five 20-second job-posting clips from one recruiter avatar cost $122.50 on Plus or $92.00 on Standard, plus $0.95 once to create the avatar, at Sume's listed rates (read 2026-10-03). That is 500 seconds of video, or $4.90 per role on Plus.
A short video at the top of a posting answers the question candidates care about, which is what the team is like, in a format they can scan. A single reusable presenter keeps the series consistent without scheduling anyone.
Build the template once
A 20-second script holds about 50 spoken words, so structure matters. Use four beats: the role and team in one sentence, the main problem the hire will solve, one thing about how the team works, and where to apply. Write one template with four slots and fill it from the job description.
Create the presenter with the avatar creation route: a prompt or structured traits for an invented recruiter, or a photo if a real recruiter has agreed to appear. A fictional presenter should be labeled as such in the posting if you do not want candidates to assume it is a real employee. Keep the handle meaningful, such as recruiter_maya.
| Item | Quantity | Rate | Cost |
|---|---|---|---|
| Create avatar | 1 | $0.95 | $0.95 |
| Standard clips | 25 x 20 s | $0.184 per second | $92.00 |
| Plus clips | 25 x 20 s | $0.245 per second | $122.50 |
| Max clips | 25 x 20 s | $0.55 per second | $275.00 |
Run the batch
Submit each role with its own Idempotency-Key built from the requisition id, so a rerun never produces a duplicate paid clip. Your plan's concurrency and queue capacity determine how many are accepted at once; 25 submissions fit within a queue on larger plans and must be sent in waves on the smallest. See Generation admission for the limits.
The balance has to cover the batch, so read GET /v1/balance first. For 25 Plus clips you want at least $122.50 free, and more if you plan retakes.
import asyncio
TEMPLATE = (
"We're hiring a {role} on the {team} team. You'll {problem}. "
"The team ships weekly and reviews work together. Apply at {url}."
)
async def main() -> None:
roles = [("data analyst", "Insights", "turn messy usage data into weekly decisions", "example.com/jobs/1")]
for role, team, problem, url in roles:
script = TEMPLATE.format(role=role, team=team, problem=problem, url=url)
print(len(script.split()), "words:", script)
asyncio.run(main())
Before you publish
Hiring content touches regulated areas in some countries. Treat the clip as part of the posting and have your usual reviewer approve the script, not the video.
- Say it is an AI presenter, in the script or caption, so candidates are not misled.
- Do not make promises about pay, visas or benefits in the video unless HR has approved the exact words.
- Re-render a clip if the posting changes; a stale video is worse than none.
- Use
captionsso the clip works muted on job boards.
Two versions of the same batch
If you want a cheaper pilot, render the first five roles on Standard, which comes to $18.40, and review them with two recruiters. If they approve the quality, run the remaining 20 on Plus for $98.00. The pilot then costs $18.40 and the rest $98.00, for a total of $117.35 including the avatar. Mixed tiers in one series are fine so long as viewers do not see the difference, so test that first.
The alternative is to approve first frames with the preview route and then render only approved roles. For hiring clips the first frame carries most of the risk, since a strange expression or a wrong background shows immediately.
What not to put in the script
Skip salary figures unless the posting itself states them, and avoid any claim about the team you cannot verify. Do not use the avatar to impersonate a real hiring manager. If the role is remote, say where candidates can work, using the exact wording from the posting. A clip that disagrees with the written description creates more questions than it answers.
Sources
Related posts
More in Use cases
- AI product angles from one photo: front, three-quarter and back views
Make extra product angles from one packshot: send a second photo of the back for printed sides, then check each view. A reference-edit request on Sume.
- AI share image for blog posts via API: 16:9 art, crop to size
Generate a text-free 16:9 hero per post on Sume from $0.025, crop it to your site's share-card size, and set the headline in code. Cost for 100 posts.
- AI sneaker product photo on a colored backdrop, from one packshot
Keep the sneaker, change the set: a prompt, a preserve list and a reference photo give colored-backdrop product stills. Request, ratio and checks on Sume.
- AI sonic logo generator: a short audio sting by prompt, then trim
Make a 2 to 4 second sonic logo with AI: write the sting into the prompt, then cut it with Timeline audio. Sume has no duration field and no fade on the cut.
Written by Sume