On this page
- The steps
- Real makeup video ads
- What a faceless brand video is
- Write a script that pairs every line with a visual
- Mix real footage with AI-generated visuals
- Add voiceover, captions and music
- What long-running makeup ads teach about openers
- Disclose, export and iterate
- Prompts to copy
- Which visual source to use for each shot
- Mistakes to avoid
- Questions
Real makeup video ads to model yours on
3,632 video ads from 365 advertisers in the Meta Ad Library, read October 5, 2026. The median one ran 22 days, and an ad still paid for after months is the closest public sign that it works. Live ads first.
- Ran under a day
- Ran 7 days
- Ran 2 days
- Ran 7 days
- Ran under a day
- Ran 4 days
- Ran under a day
+3,620 more makeup video ads
Hooks, landing pages and run length for each
- Ran under a day
- Ran 6 days
- Ran 6 days
- Ran 4 days
Every creative, with its copy, format and run length
The openers that ran longest
40% OFF + FREE Tint โ Lashes Youโll LOVE!
Luxe Cosmetics ยท ran 504 daysChat With Us
Isabella Garcia International ยท ran 455 daysRealize Sonhos, Seja Consultora
Natura ยท running 434 daysBrushes Are Back in Stock
Morphe ยท ran 429 days
What a faceless brand video is
A faceless brand video has no presenter, no avatar and no talking head. The viewer sees your product, your hands, your screen or a designed scene, and hears a voice. That is the whole format. It works because the demonstration carries the message, and it removes the hardest parts of production: casting, lighting a person and coaxing a natural performance.
AI fits into three jobs. It writes the script and shot list. It generates visuals you cannot easily film, such as clean backgrounds, abstract textures and transitions. It can also voice the narration if you would rather not record yourself. It should not replace footage of the real product doing the real thing, and it should never invent claims, testimonials or results.
Write a script that pairs every line with a visual
Give your AI writing tool the brand, the product, the audience, one benefit and the call to action. Ask for a table with four columns: timestamp, voiceover, visuals and on-screen text. Tell it to open with a hook in the first three seconds, and tell it not to invent product claims, testimonials or results. A 30-second video holds roughly 70 to 80 spoken words, so cap the voiceover there.
Use a simple timing structure. Seconds 0 to 3: the hook, spoken and on screen. Seconds 3 to 10: the problem or the result the viewer wants. Seconds 10 to 24: the demonstration in three or four shots of 2 to 5 seconds each. Seconds 24 to 30: brand name and one call to action. Every voiceover line needs a visual you can actually produce.
A desk organizer brand might open on a cluttered desk with the text "Still digging for your charger?", cut to hands putting each cable in its place, and end on the tidy desk with "Find your desk setup." Review the draft line by line and delete anything you cannot prove with footage.
- Ask for one benefit, not five
- Cap the voiceover at about 75 words
- Put on-screen text in the hook
- Remove any claim you cannot show
Want makeup video ads like these for your product?
The makeup video ads that kept running, read from the ad libraries. Every change waits for your approval.
Mix real footage with AI-generated visuals
Film the product in use on your phone, held vertically, near a window for soft light. Prop the phone on a tripod or a stack of books. Good faceless shots include hands applying, pouring, unboxing or arranging, plus close-ups of texture and finish. For software, record your screen in a phone-shaped layout so the interface is readable without zooming.
Use AI for what you cannot film cheaply: a styled backdrop, a macro texture, a scene-setting establishing shot or a transition between clips. Where your tool supports it, upload a clean product photo as a reference image and animate it. Keep each clip to 3 to 5 seconds with one simple camera move. Avoid generated hands and generated text, which models often distort.
The rule is AI for atmosphere, real footage for proof. If the viewer is judging what the product does, a generated clip misleads them. Use generated clips before and after the demonstration, never in place of it.
Add voiceover, captions and music
To record your own voice, use your phone in a room with soft furnishings, such as a closet of clothes, with the mic a hand's width from your mouth. Record one line per take so you can redo single lines. If you use a licensed AI voice, confirm the license covers commercial ads, choose a natural pace and listen for mispronounced product names.
Build the edit in Canva, CapCut or Adobe Express. Drop the voice track in first, then cut visuals to it. Burn in captions of 3 to 5 words at a time, large enough to read on a phone, and keep them clear of the bottom edge and the right side where platform buttons sit. Proofread the brand name and every number.
Add music last and keep it clearly below the voice. If you have to strain to hear the narration, the music is too loud.
What long-running makeup ads teach about openers
The Sprites ads library shows real ads from a category. In makeup, the Meta Ad Library holds 3632 video ads from 361 advertisers, and the median run is 21 days. Against that, Rise Luxus ran "Buy 1 Get 1 FREE Today!" for 372 days. Sheryl Davis ran "Flawless Skin In Seconds" for 294 days. Vibriance ran "Finally, Tinted SPF Made For Women 50+" for 280 days. Wonderskin ran "The Blush That Becomes Uniquely Yours" for 185 days.
The pattern is a single line that states an offer, a fast result or a named audience. None of these openers needs a face to land. Each works as on-screen text over product footage, which is exactly the faceless format. Copy the structure, not the words: lead with the deal, the result with a time frame, or the person it is for.
For your video, put that line on screen in the first three seconds over a close-up of the real product. Then match the closing line to the button. In this category the most common buttons are Shop now, Learn more and See details. For more category patterns, see makeup video ads.
Disclose, export and iterate
Platforms have rules for synthetic media. YouTube asks creators to disclose realistic altered or synthetic content, and TikTok has labels for AI-generated content. Stylized backgrounds are rarely the issue. A generated clip that looks like a real person, event or product result is. Read each platform's current policy before you publish, and check whether a realistic synthetic voice counts.
Export as MP4 at 9:16, 1080 by 1920. Confirm placement details in the Meta ad specs and keep captions inside the safe zones. Then watch the file on your phone twice, once with sound and once muted. If the muted version does not explain the offer, add or sharpen on-screen text.
To iterate, keep the body fixed and change only the first three seconds. Make three hooks: an offer, a result and an audience callout. Run them side by side and keep the winner. For a broader process, read the video ads guide. If you would rather not produce it yourself, 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: product reveal
Vertical 9:16, 4 seconds. Slow push-in on a [product] resting on a clean marble surface beside a window, soft morning light, shallow depth of field. Subtle dust particles in the air. No people, no faces, no hands, no text, no logos. Calm, premium feel, steady camera, natural colors, sharp focus on the product.
Hook: problem scene
Vertical 9:16, 3 seconds. Top-down view of a cluttered wooden desk with tangled cables, loose papers and a coffee mug, warm indoor light. Camera drifts slowly to the right. No people, no faces, no readable text, no logos. Realistic style, slightly messy, handheld feel, muted colors that suggest frustration.
Demo bridge: texture macro
Vertical 9:16, 4 seconds. Extreme macro of a smooth cream texture being swirled slowly on a glass surface, soft studio lighting, pastel background. Camera locked off, slow motion, glossy highlights. No people, no hands, no faces, no text, no logos. Clean, satisfying, high detail.
Transition: abstract wipe
Vertical 9:16, 2 seconds. Soft colored paint ribbons sweep across the frame from left to right, revealing a plain bright background. Smooth fluid motion, pastel palette matching [brand color], no objects, no people, no text, no logos. Designed to sit between two real product clips as a quick transition.
Closer: clean end background
Vertical 9:16, 5 seconds. A tidy minimal tabletop scene with a soft gradient wall in [brand color], a small potted plant on the left and open empty space in the center for text. Gentle light shift, locked camera. No people, no faces, no text, no logos. Calm and finished, leaves room for a call to action.
Which visual source to use for each shot
| Shot type | Best source | Use it for | Watch out for |
|---|---|---|---|
| Product in use | Your phone footage | Proof of what the product does | Shaky framing and dim light |
| Screen recording | Your own screen capture | Software walkthroughs | Text too small to read on a phone |
| Background or backdrop | AI-generated clip | Scene setting before and after the demo | Warped objects and odd text |
| Texture or macro | AI-generated clip or real close-up | Bridging between demo shots | Showing a finish the product cannot deliver |
| Transition | AI-generated clip or editor effect | Cutting between sections | Long clips that slow the pace |
Mistakes to avoid
- Opening on a logo animation instead of the viewer's problem or wanted result.
- Using AI-generated footage to show what the product does, which misleads viewers.
- Letting the AI script invent testimonials, claims or results you cannot prove.
- Writing a voiceover longer than 75 words for a 30-second video, which forces rushed pacing.
- Placing captions where platform buttons cover them, or skipping the muted watch test.
- Letting music compete with the voice instead of sitting clearly below it.
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 โQuestions
Can I make a faceless brand video without an AI avatar?
Yes, and it is usually the better choice. Combine product footage or screen recordings with a voiceover and captions. Use AI for scripting and for background visuals, not for a fake presenter.
What is the best length for a faceless brand video?
Thirty seconds is a strong starting point because it fits one benefit and one call to action. Open with a hook in the first three seconds. Shorter is fine if the message is already clear.
Should I use my own voice or an AI voice?
Either works. Your own voice recorded off camera is simple and avoids licensing questions. If you use an AI voice, make sure it is licensed for commercial use and check that it pronounces your product name correctly.
Do I have to label AI-generated visuals in my video?
It depends on the platform and how realistic the content looks. YouTube asks for disclosure of realistic altered or synthetic content, and TikTok has labels for AI-generated content. Read the current policy before you post.
What aspect ratio should a faceless video be?
Use 9:16 vertical for short-form placements, exported at 1080 by 1920. Keep captions and key visuals away from the edges. Check the exact placement rules for each platform before upload.
How do I know if my faceless video works?
Watch it on your phone with sound, then muted. You should be able to say what is offered and what to do next both times. After publishing, test different first three seconds against the same body.