A Skool classroom is a feed still, not a twin teacher

Skool’s listing is the classroom plus community feed. Overlay the module names. Do not generate a synthetic instructor for the muted ad.

A Skool clip that converts is the object a member already sees: classroom rows, community posts, the module list. That screen is a still. Motion is optional. A constructed teacher pointing at a fake whiteboard is a different object, and it is not the listing. Typeset the feed. Burn the module names. Keep the twin off the muted file.

The listing is the classroom plus community feed

Skool is not a LMS screenshot and not a campus tour. The product is the classroom attached to a community: lessons in one column, posts in the other, modules a member can name. Capture the real community you would ship, or typeset a still that holds those two panes. Crop other people’s avatars if a test account leaked them. Crop DMs. The still is a document. Treat it like one.

Do not prompt a video model to “show a vibrant online classroom with a teacher on camera.” It will invent a cousin platform, a cousin syllabus, a cousin face. A prospect who lands in Skool and sees a different module list is not looking at creative. They are looking at a lie.

Online course creators convert with a real lesson excerpt, not a testimonial reel. That page is teaching quality. This page is the Skool listing around the teaching. Do not twin a free lesson as a second feed still.

Membership site owners own churn after the join: public proof versus member library. Skool’s classroom-plus-feed is one SKU inside that world. Do not rewrite the membership calendar. A K–12 campus tour is a different audience. A Skool community is not a school.

If you already have a screenshot of the live classroom, that is the honest still. Generate only when you need a clean crop the live UI will not give you.

Qwen for the feed still

When the still has to be a classroom-plus-feed you do not want to crop from a messy live tab, use an image row that can hold letters.

Qwen Image 3 Text to Image is that still: 4 credits, up to 2K, text rendering in English and Chinese, prompt rewriting on by default. Quote every string a member will proof. Community name. Module titles as you will actually publish them. Open it from text-to-image. Scout until the hierarchy reads. Keep at 2K when someone will zoom.

Quote the strings. “A stylish Skool classroom” is how you get a plausible sidebar you cannot ship. Proof at 100%. One wrong character in the community name is a regenerate. Qwen is not Skool. It will not enroll a member. It typesets a picture of a feed.

Do not ask Qwen for a photoreal instructor in a ring light holding a laptop that says “Module 1.” That is a host. The SKU is the feed. If the type will be overlaid later, generate the plate with empty module slots rather than hoping the model spells every title.

Prompt rewriting is on by default. If the model “helps” your module names into synonyms, the listing is wrong. Quote the titles in marks. Proof again.

Overlay the module names

The module list is a string you already set in Skool. Generated type inside a video is a misspelled lesson you cannot edit without a new generate. Burn the string after picture lock.

Add a text overlay to a video burns one authored line for the whole clip: the community name, the hook, the CTA, a single module stamp. You write the line. Position is top, center, or bottom. Font size, color, and an optional background block are controllable. It does not transcribe. A mute feed still has nothing to listen to.

If you needed three different lines at three timestamps — “classroom,” then “community,” then the module name — that is timed overlays, not one persistent badge. Picture lock first. Overlay before you reframe to 9:16 and you will pay twice: the line that sat clean on a desktop feed sits in the chrome after a vertical crop.

Do not put a made-up earnings line on the overlay. Sell the classroom, not an implied income.

A twin teacher is the muted-ad mistake

Muted autoplay is how this ad ships. A constructed instructor talking into silence is a face you do not have and a transcript you cannot run. Add captions to a video transcribes speech. The feed still has none. Running a caption pass hoping “Module 3” appears is a spend on silence.

Do not:

• Text-to-video a studio with a glowing “Join Skool” button and a twin of yourself.
• Generate a teacher pointing at a whiteboard the model invented.
• Bake the module list into the generate so a curriculum change forces a new take.

Change the module with an overlay. Change the community with a new screenshot. Keep the host off this plate. If a human must speak, it is a real named instructor on a different day, on a file that is not pretending to be the listing — and then you caption that take.

Pause the keeper. Can you read the community, and is there no invented mouth? Yes: overlay the modules. No: you used a presenter row as a classroom.

FAQ

Can I generate a prettier classroom instead of capturing Skool?

You can typeset a mock on Qwen if you quote the real strings and proof them. You cannot invent a community name, a module list, or a teacher. The live screenshot is the honest SKU. Overlay the production titles if the mock is only a crop.

Is overlay the same as auto-captions on a voiceover?

No. Overlay burns a line you wrote. Auto-captions transcribe speech. A mute feed has nothing to transcribe. If you later record a real instructor, caption that file separately.

Why not put the module names in the video prompt so the model draws the sidebar?

Because a title is type, and type inside a generate is a typesetting problem you cannot edit. Overlay the names after picture lock. A curriculum change then costs a burn, not a new cinematic.

Can the online-courses /for playbook replace this still?

No. Online courses is a real lesson excerpt. A Skool listing is the classroom-plus-community feed. A generated teacher is not that excerpt and not this listing. Overlay the modules. Keep the twin off the muted ad.