Akool max video length: 15 to 60 minutes by plan vs Sume's 60 s

Akool's pricing page lists a maximum video length of 15 minutes on Starter, 30 on Pro, 45 on Pro Max and 60 on Business. A single Sume Avatar 1.0 talking-video job accepts an estimated 4 to 60 seconds, so anything longer is several jobs that you plan and join yourself.

The two numbers are not the same kind of limit. Akool's is a ceiling on one video inside a credit plan. Sume's is the window for one avatar render, billed per second.

What Akool lists

Read on 2026-10-07, the Akool page shows these plan tiers and ceilings. Prices are the per-seat figures shown for yearly billing.

What the same length costs on Sume

Sume's avatar video guide says to shorten longer scripts or split them into several jobs. The catalog rates are $0.184 (standard), $0.245 (plus) and $0.55 (max) per second without a product image.

Fifteen minutes is 900 seconds, or fifteen 60-second jobs: $165.60 on standard, $220.50 on plus, $495.00 on max. Sixty minutes is 3,600 seconds: $662.40, $882.00 and $1,980.00. Those are list-rate arithmetic, not a quote, and they exclude any step that joins the clips.

How to read the comparison

Akool pools credits across features, so a 60-minute cap does not mean 60 minutes are included in a plan. I read no per-minute avatar price on that page. The useful question is how long your real videos are.

If your videos run under a minute, such as ads, hooks and short replies, the cap is irrelevant and you pay for seconds. If you make 20-minute lessons, plan the script as scenes of 60 seconds or less and budget the join. Sume does not render a single 15-minute avatar take.

Splitting a long script

Sume lets one request carry up to 20 ordered scenes in video_inputs, but the total planned duration must still fall in the 4 to 60 second window. So a 10-minute talk is at least ten requests. Give each request the same avatar_handle and a consistent scene prompt, and keep an Idempotency-Key per request so a retry does not bill twice.

The cost scales linearly, which makes it easy to forecast: seconds times the tier rate, per job. That is the main difference from a plan with a ceiling, where the marginal minute is free until you hit the cap.

A short checklist

• Count your script's seconds, not its pages: the API estimates duration from the text.
• Multiply by the tier rate, then add $0.01 to $0.03 a second if you pass a product image.
• Keep one avatar handle across jobs so every clip shows the same face.