Faceless TikTok for Your App: What to Put on Screen
Search "faceless TikTok ideas" and you get motivation quotes over drone footage, Reddit stories read by a robot voice, and "top 10 facts" channels. None of that sells an app. The viewer never sees the product, never sees it solve anything, and never learns the name.
Faceless TikTok for my app has to do one job: put the product on screen, show the problem it fixes, and do it in a shape TikTok will distribute. Below are the four formats that work for software, what goes in each scene, and the specs to export to. Nothing here needs a camera or a creator invoice.
The rule all four formats share
The screen is the product. Not stock footage of a laptop, not an abstract gradient. Real screenshots, real product photos, the actual icon. TikTok without showing my face still means showing something a viewer can recognise next time they see it in the App Store.
Three constraints shape every scene:
- Export at 1080 x 1920, 9:16. Square or landscape uploads get pillarboxed inside the vertical player (recurpost).
- Keep text away from the edges. The right-hand column holds the like and comment buttons, the bottom holds the caption. Third-party guides disagree on exact margins (roughly 130 to 480 px at the bottom, 130 to 150 px at the sides), so the safe move is to keep text inside a centred box of about 880 x 1420 px (clickanalytic).
- The first scene is the hook. On-screen text of five to eight words, readable with sound off, stating the problem or the payoff (Skeepers).
Format 1: the screenshot scene sequence
This is the workhorse. Five to eight scenes, each one a real screenshot or product photo, two to three seconds each, with one line of on-screen text and a voiceover reading something close to it. Add gentle motion (a slow zoom or pan) so the video carries watch time, which is the signal TikTok rewards and which a static photo post cannot earn.
A worked example for a recipe app:
- Hook: photo of a fridge with random ingredients. Text: "Four things in the fridge, no idea what to cook."
- Screenshot of the ingredient input screen. Text: "Type what you have."
- Screenshot of the results list. Text: "It gives you three dinners you can actually make."
- Screenshot of one recipe, timer visible. Text: "Steps with a timer, no scrolling past a life story."
- Photo of the finished plate. Text: "Done in 25 minutes."
- App icon on a plain background with the name spoken once.
Notice what the scenes avoid: no feature grid, no "powered by AI" slide, no pricing. One problem, one path through the app, one result.
You can build this with free tools. Number your screenshots and run ffmpeg:
# 6 screenshots, 2.5 s each, filled to 9:16, voiceover mixed in
ffmpeg -framerate 0.4 -i scene%d.png -i voiceover.m4a \
-vf "scale=1080:1920:force_original_aspect_ratio=increase,crop=1080:1920,fps=30" \
-c:v libx264 -pix_fmt yuv420p -c:a aac -shortest out.mp4
The scale-then-crop filter fills the frame with no black bars (Editframe). Put the on-screen text into the PNGs themselves, or add it in TikTok's editor after upload. Phone screenshots are already 9:16, so they pass through untouched. Website screenshots need cropping to a phone-shaped region first, or they will be zoomed in beyond recognition.
Social Studio's photo video format is this exact structure: a sequence of scenes with motion, on-screen text, music and a voiceover, built from the product photos and screenshots you upload, so the product in every scene is the real one. See the live examples at studioposting.com/samples.
Format 2: the before/after problem reveal
Two halves. The first shows the mess the user lives with today. The second shows the same situation inside your app. It is the strongest format for try-on, beauty, fitness and productivity apps, because the "before" is a feeling the viewer already has.
Before scenes do not need the app at all: a photo of a cluttered notes app, a screenshot of a 47-tab browser, a mirror selfie with three lipstick shades drawn on the back of a hand. After scenes are your screenshots. Cut the two halves with a hard transition and change the music energy on the cut.
The hook text names the before: "My workout log was a Notes file with 400 lines." The last scene names the after and the app, once.
One thing to avoid: do not fake the before with an edited screenshot of a competitor. A photo of your own old spreadsheet is more believable and carries no risk.
Format 3: "three things this app does"
A numbered list in video form. Hook: "Three things [app] does that [category] apps usually don't." Then three scenes, each a screenshot with a number and a short claim, then the icon.
This format exists for discovery. Viewers who search TikTok for "best budgeting app" or "AI photo app" land on list content, and a three-item list from the maker reads as information rather than an ad. Keep each claim checkable from the screenshot behind it. "Syncs with Apple Health" with the permissions sheet on screen is fine. "Loved by thousands" over a gradient is not.
Rotate the three. The same app can produce ten of these videos by changing which three features lead, and each one is a fresh entry into the same product.
Format 4: a consistent persona who is not you
Some founders want a face on screen for the trust it gives, without it being their own. The faceless solution that works is a consistent generated persona: the same face, hair and style in every scene, across every post, so the account feels like one presenter rather than a stock-photo lottery.
Two rules make this safe on TikTok. First, the persona must be visibly consistent; swapping faces between posts reads as spam. Second, label it. TikTok's Community Guidelines require creators to label AI-generated or significantly edited content that shows realistic-looking scenes or people, and unlabeled content can be removed, restricted or labelled by TikTok (Community Guidelines). The label is a toggle under the post's options in the app, and apps that publish through the Content Posting API set the same flag with the is_aigc field (TikTok for Developers). We wrote up what the label does and does not affect in the AI-generated label post.
Social Studio's persona option works this way: faceless by default, and if you switch the persona on, the generated face stays the same across scenes. Every post ships with the AI-generated-content disclosure regardless.
Why carousels stay a small share
TikTok photo posts (carousels) are tempting for a faceless founder because they are the cheapest thing to make: up to 35 images, one caption, pick a sound (Content Posting API photo post). The problem is that they carry no watch time. A viewer swipes through in four seconds or does not swipe at all, and for most accounts that means far less reach than a video built from the same images.
Use carousels for the content that genuinely reads better as slides: a step-by-step setup guide, a "save this" reference list. Keep them to a small share of the mix and put the screenshot scene sequence first. The longer comparison, with the cases where carousels win, is in carousel vs video.
How to run this as a solo founder
TikTok content for a solo founder fails for one of two reasons: making each video takes an evening, or the account gets no views in week one and the founder stops.
For the first: build a scene bank once. Take 20 to 30 clean screenshots with a demo account (no personal data, consistent light mode or dark mode, status bar tidy). Take ten real photos of the thing your app is about: the fridge, the gym bag, the guitar. Every video for the next month is a reshuffle of that bank with new text and a new voice line. Two screenshot sequences, one before/after and one list per week is a realistic cadence from a bank that size.
For the second: a new account's first posts can sit at near-zero views, and that is about the account, not your hook. Judge a format over a week of posts, not one upload. The diagnosis and the fixes are in 0 views on a new account.
If you would rather not do the assembly at all, Social Studio plans the ideas on a calendar, builds each post in these formats from your uploaded product photos with an original music track, a voiceover from a catalog of 200+ voices, and a written caption and hashtags, then holds it until you press Post. Details for app makers are at studioposting.com/for/apps.
A one-screen checklist
- Real screenshots or product photos in every scene, never stock.
- 1080 x 1920, text inside the centre safe box.
- Hook text of five to eight words on scene one.
- Two to three seconds per scene with motion, voiceover close to the text.
- App name spoken and shown once, at the end.
- Persona on? Label it as AI-generated.
- Carousels only for slide-native content, and rarely.