On this page
- The steps
- Real mobile app video ads
- The 15 second structure for app install ads
- What long-running app ads open with
- Generate the hook in Dreamina
- Record the real app demo
- Assemble and add text in your editor
- Test three hooks, then scale
- Prompts to copy
- What to generate and what to record
- Mistakes to avoid
- Questions
Real mobile app video ads to model yours on
3,317 video ads from 298 advertisers in the Meta Ad Library, read October 5, 2026. The median one ran 38 days, and an ad still paid for after months is the closest public sign that it works. Live ads first.
- Ran 1 day
- Ran 2 days
- Ran under a day
- Ran under a day
- Ran 1 day
- Ran under a day
- Ran 2 days
+3,305 more mobile app video ads
Hooks, landing pages and run length for each
- Ran 1 day
- Ran under a day
- Ran 4 days
- Ran under a day
Every creative, with its copy, format and run length
The openers that ran longest
ΰ€ͺΰ€Ήΰ€²ΰ€Ύ ΰ€ΰ₯ΰ€² @Rs 1/- π
InstaAstro Β· running 1,395 daysCuentos en 23 idiomas para ti!
Beelinguapp Β· running 756 daysClassic merge puzzle game
Gossip Harbor Β· running 748 daysJoin Local Groups & Make New Friends
Meetup Β· running 590 days
The 15 second structure for app install ads
Seedance 2.5 is good at believable motion and lifestyle footage. It is not a screenshot tool. If you ask a video model to render your app interface, you risk garbled labels and features your app does not have. So split the job. Generated video carries the hook and the payoff. A real screen recording carries the proof. Your editor carries every word on screen.
Here is the timeline. From 0 to 3 seconds, show a problem your audience recognizes at once. From 3 to 10 seconds, show one action in the app and its result. From 10 to 12 seconds, hold on the payoff. From 12 to 15 seconds, show the app name, logo and one call to action.
This structure also makes iteration cheap. The hook is its own clip, so you can swap it without remaking the demo. Seedance 2.5 can run up to 30 seconds in one generation, but for app ads you gain more from three short, separate hooks than from one long clip. For more formats and examples, see mobile app video ads.
- Generate: hook, lifestyle shots, reaction shots
- Record: the app demo
- Build in your editor: captions, logo, end card
What long-running app ads open with
The Meta Ad Library shows 3,317 video ads from 285 mobile app advertisers, with a median run of 38 days. The openers that ran far longer share a pattern. Meetup ran "Join Local Groups & Make New Friends" for 589 days with a Download button. Opal ran "A standing desk you can trust" for 558 days. Spotify ran "Give every step a soundtrack." for 331 days with Install now.
Each line names an outcome in plain words. Meetup says what you get: friends and local groups. Spotify ties the product to a moment: walking. Whatnot ran "Yup, you can buy new outfits without breaking the bank." for 386 days, which opens on a recognizable worry about cost and answers it. None of them lead with a feature list.
Translate that into Seedance prompts. Generate the moment the line describes: a person arriving at a group meetup, a walk with no one in frame, a closet full of clothes. Then match the button to the goal. The most common buttons in this category are Install now, Learn more, Download and Sign up. The Sprites ads library shows real ads from a category, so you can scan openers before you write yours.
Want mobile app video ads like these for your product?
The mobile app video ads that kept running, read from the ad libraries. Every change waits for your approval.
Generate the hook in Dreamina
Open Dreamina, then AI Video, then Seedance 2.5. Choose 9:16 for a vertical ad. If you want a visual or motion reference, use Omni reference and upload a still of the setting or a clip of the camera move you want. Do not upload anything with a logo you do not own.
Write the prompt like a shot brief. State the aspect ratio, who is doing what, the camera, the lighting and one believable emotion. Tell the model to start with motion already in the first frame, so the clip does not open on a static pose. Add a line that keeps the upper third clear for a caption. Say no on-screen text, logos or dialogue.
Generate the hook as its own short clip, even though the model can go longer. You will trim it to about 3 seconds in your editor. Change one thing per run: the setting, the reaction or the camera. That way you know what caused each difference when you compare results.
- One continuous handheld shot reads like a real phone video
- Keep the subject in the lower two thirds
- Avoid visible phone screens in the generated clip
- Save the best three takes, not just one
Record the real app demo
The demo is the only part viewers will believe is your product, so record it for real. Use a clean test account with sample data that fits the hook. Turn on Do Not Disturb so no notification slides in. Record one action and the result it produces, such as selecting ingredients and seeing a recipe suggestion.
Crop in during editing. A full phone screen shrunk into a vertical frame makes the tap hard to see. Zoom to the area where the action happens, then follow the result. Keep the demo near 7 seconds. If it needs more, you are showing two features, and the ad should pick one.
Match the demo to the hook. If the hook shows someone staring into a fridge, the first demo frame should be the app step that fixes that exact problem. A jump from a story to an unrelated screen breaks the ad in the first second after the cut.
Assemble and add text in your editor
Bring the hook, demo and payoff into CapCut or your editor of choice. Trim the hook to 3 seconds. Add a short caption in the clear upper third, for example 'Nothing to cook?' Cut to the demo on the first moment of motion, not after a pause.
Build the end card as a plain graphic: app name, logo, one benefit line and one call to action such as 'Download [App Name].' Editors render text sharply and spell it correctly, which a video model may not. Hold the card for about 3 seconds.
Before you export, check placement rules for each channel. Keep captions and the end card out of the areas the platform covers with its own interface. The Meta ad specs and TikTok ad specs pages list the safe zones and file limits. Then watch the ad on a phone with the sound off. If you cannot tell what the app does, simplify the demo before adding effects.
Test three hooks, then scale
Make three versions that change only the opening. Version one shows a relatable problem. Version two shows the finished result. Version three states a direct benefit in the caption over a calm lifestyle shot. Keep the demo and end card identical in all three.
Run them with the same audience and the same button. Compare which opening holds viewers into the demo, then keep the winner and write three new variations of it. Long-running openers in this category, like Opal at 558 days, stayed on one clear promise instead of changing the message each week.
Keep a simple log: the prompt, the reference used, the caption and the result. When you find a hook that works, reuse its structure with a new setting or character. If you want a team to handle this routine for you, Director by Sprites is a full creative team for your startup. Every Monday you get finished, on-brand video ads, reviewed by a human.
Prompts to copy
Paste each one as written, then change the product, the setting and the light to match your own footage.
Hook: relatable problem (meal planning)
Vertical 9:16, natural smartphone-shot ad. An adult opens their fridge after work, looks at a few random ingredients, and pauses, unsure what to cook. Start with the fridge door already moving in the first frame. One continuous handheld shot, eye-level camera, ordinary kitchen lighting, believable reaction. Keep the upper third clear for a caption. No visible phone screen, on-screen text, logos, or dialogue.
Hook: finished result (meal planning)
Vertical 9:16, natural smartphone-shot ad. A freshly cooked pasta dinner is set on a kitchen table as hands place a fork beside the plate. Start mid-motion with the plate already landing. Warm evening light, slight handheld movement, shallow depth of field. Keep the upper third of the frame empty for a caption. No phone screens, on-screen text, logos, or dialogue.
Hook: direct benefit (budgeting)
Vertical 9:16, casual phone-style video. A person at a cafe table looks relieved while checking a paper receipt, then smiles and slides it into a wallet. Start with the receipt already in hand. Soft daylight from a window, steady handheld camera, medium close-up, natural expression. Leave the upper third clear for a caption. No visible phone screen, on-screen text, logos, or dialogue.
Lifestyle shot: music or fitness app
Vertical 9:16, tracking shot following a runner from behind along a city park path at golden hour. Start already in motion, with leaves moving in the wind. Smooth handheld camera, warm light, relaxed stride, earbuds visible but no brand marks. Keep the sky area in the upper third clear for a caption. No phone screen, on-screen text, logos, or dialogue.
Payoff: reaction shot
Vertical 9:16, close-up of a person at a kitchen counter glancing down at their phone out of frame, then breaking into a relieved smile and nodding. Begin mid-expression. Natural window light, shallow depth of field, slight handheld motion. Keep the upper third clear. Do not show the phone screen. No on-screen text, logos, or dialogue.
Omni reference: motion match
Vertical 9:16 ad. Use the uploaded motion reference for the camera move: a slow push-in from a wide room view to a close-up of a person sitting on a couch. Match the pacing and framing of the reference. Soft living room lighting, relaxed mood, subject looks at something off camera and exhales. No visible phone screen, on-screen text, logos, or dialogue.
What to generate and what to record
| Segment | Time | Source | Why |
|---|---|---|---|
| Hook | 0 to 3 sec | Seedance 2.5 in Dreamina | Fast, flexible, easy to swap for testing |
| App demo | 3 to 10 sec | Real screen recording | Shows true UI and real features |
| Payoff | 10 to 12 sec | Hold on result or generated reaction | Gives the viewer the emotional close |
| End card | 12 to 15 sec | Your editor | Keeps the app name, logo and CTA sharp |
| Captions | Throughout | Your editor | Accurate spelling and placement |
Mistakes to avoid
- Asking Seedance 2.5 to render your app interface, which risks garbled labels and invented features.
- Generating the whole 15 second ad in one clip, which makes it hard to swap the hook.
- Opening on a static pose instead of motion already happening in the first frame.
- Leaving no clear space in the upper third for a caption.
- Showing the full phone screen in the demo so the tap and result are too small to read.
- Testing three hooks with three different demos, which makes the comparison meaningless.
Rather have the videos made for you?
Director by Sprites is a full creative team for your startup. Every Monday you get finished, on-brand video ads, reviewed by a human. You never write a prompt.
Apply to Director by Sprites βQuestions
Can Seedance 2.5 make a full mobile app ad on its own?
It can make the hook, lifestyle and reaction shots well. For the demo, record your real app, because generated screens can show wrong labels or features you do not have. Add the end card in your editor.
Where do I use Seedance 2.5?
Open Dreamina, go to AI Video, and select Seedance 2.5. Choose 9:16 for vertical ads and use Omni reference if you want to add a visual or motion reference.
How long can a Seedance 2.5 video be?
Seedance 2.5 supports up to 30 seconds in one generation. For app ads, generate the 3 second hook separately so you can swap it without remaking the demo.
What aspect ratio should a mobile app video ad be?
Use 9:16 for vertical placements on feeds and short video. Generate in 9:16 from the start so you do not crop the clip later. Check the channel's spec page for safe zones and file limits.
How do I add text to a Seedance 2.5 ad?
Add it in your editor, not in the prompt. Tell the model to leave the upper third clear and include no on-screen text, then place the caption and the call to action as graphics.
How many hook variations should I test first?
Start with three: a relatable problem, the finished result and a direct benefit. Change only the opening and keep the demo and end card identical. Then build new variations around the winner.