01
Step 1: Pick a tidy, legible scene
Tutorials read best with a clean background and uncluttered surface so the step is obvious. Clutter competes with the instruction and also gives the model more places to slip.
Use case · AI tutorial video generator
A tutorial needs a clear instructor and legible hands-on steps. We generated a demonstration frame to see how the hands-visible composition holds.
socialAF research pipeline·Generated June 1, 2026

What we ran
A clean tutorial frame with the saved character demonstrating the product, hands visible mid-gesture, rendered at 2 credits.
We generated an instructional frame of the character showing the product with hands mid-gesture on a tidy surface. The instructor matches the set. Hands-visible demo frames are harder than talking-head frames, so a tidy background and a single clear gesture help the model.
How to reproduce·generate_image with the saved character plus product reference. Prompt gist: clean tutorial demo, hands visible mid-gesture, tidy surface, 4:5.
01
Tutorials read best with a clean background and uncluttered surface so the step is obvious. Clutter competes with the instruction and also gives the model more places to slip.
02
Ask for a single clear gesture, hands visible, rather than a complex two-handed action. We generated the demo at 2 credits with the character showing the product mid-gesture.
One legible beat per frame assembles into a clean step sequence.
03
Check hands at full size and keep the instructor consistent across steps by reusing the saved character. Generate a couple per step and select.
If a gesture looks unnatural, simplify it. A pointing hand is more reliable than a precise grip.
04
A tutorial benefits from narration. Pair the instructor still with a lip-synced or voiced track so each step has a clear spoken instruction, all from the same character.
FAQ
Yes, when you reuse the saved character. Every step frame shows the same instructor.
Use a tidy scene, ask for one simple gesture, inspect at full size, and pick from a couple of generations.
Yes. Pair the instructor still with a lip-synced or voiced track for spoken steps.
2 credits on the standard character model.
Use a tidy background, show one gesture per frame, and reuse the same instructor across steps so the sequence reads as one lesson.
Yes. Pair the instructor still with a lip-synced or voiced track so each step has spoken narration from the same character.
Ready
Join creators using socialAF to bring their characters to life. One subscription, every model, no shoot required.
Start todayUnboxing lives on hands and the reveal. We generated the unboxing frame with a referenced product. Here is the step-by-step and what to watch.
Review content compounds when the reviewer is the same person every time. We generated a review frame with a saved character. Here is the run.
We turned one character still into a lip-synced talking-head ad in about 40 seconds for 5 credits. Here is the exact two-step run.