A $10 avatar video pilot: 49 seconds on Sume vs free plan limits

Short answer

Ten dollars on Sume pays for one avatar at $0.95 and then 49 seconds of Avatar 1.0 video at standard, 36 seconds at plus or 16 seconds at max. Two vendors list free tiers for the same kind of test: HeyGen Free with 3 videos a month at 1 minute, and Synthesia Basic with 10 minutes a month and a watermark. A paid test costs a little, but it is a way to try the API route end to end.

The pilot arithmetic

Start from $10.00 and subtract the $0.95 avatar, leaving $9.05. Divide by each rate and round down.

The free tiers as listed

Both free tiers are limited in ways that matter for a pilot.

A pilot plan that fits $10

Make the cheapest valid tests first. A 4-second clip at standard costs 4 x $0.184 = $0.74. Use it to check the whole loop: create the avatar, poll the job, generate the video, fetch the result, and open the media URL. Then use a preview to look at the first frame of a longer clip before you render it.

• Avatar creation: $0.95.
• One 4-second smoke test at standard: $0.74.
• One 15-second clip at standard: $2.76.
• Remaining after those: $10.00 - $0.95 - $0.74 - $2.76 = $5.55.

What the pilot should answer

Before you spend more, decide whether the look is acceptable at 720p, whether the voice and pacing suit your brand, whether the job flow fits your system, and what disclosure you will use. A free plan tells you about a studio interface; a paid API pilot tells you about integration. Do both if you can. Keep the numbers: a clip you like at standard may be worth rendering again at plus, and the preview lets you change the final tier without making a new preview.

If you already have an avatar, the whole $10 goes to video: $10.00 / $0.184 = 54 seconds at standard, 40 seconds at plus ($10.00 / $0.245 = 40.8) and 18 seconds at max ($10.00 / $0.55 = 18.2). That is roughly three 15-second spots at standard for $8.28, with $1.72 left over. Treat these as upper bounds, since a script that Sume estimates as longer than you planned costs more seconds. Keep the pilot honest by writing the scripts first, estimating their spoken length, and only then choosing a tier.

Finally, write down what a pass looks like before you start: for example, the presenter looks acceptable at 720p, the clip is delivered as a public media URL you can open, and the disclosure line is readable. A clear bar stops a $10 pilot from turning into an open-ended experiment.