Upload a clip you already have, a podcast cut, an interview, a talking-head take. GhostShorts puts it on the top half, a gameplay clip from the library on the bottom half, transcribes the audio and prints word-synced captions across the seam, then renders a 1080x1920 MP4. Renders take about a minute. There is no free step on this template: it needs your footage first.
Upload a clip10 seconds
Read from the clip itself, not from a saved project.
Four steps, all of them in the Split Screen wizard.
In the Clips section add the video you want people to watch. This is the half that carries the audio.
Choose gameplay from the library or upload your own, then set the split layout and which half each clip sits in.
Press Generate AI captions. Your audio is transcribed and every word is timed, then you can edit the words and the style in the Captions section.
Press Generate video. The render takes about a minute and lands in your library as a 1080x1920 MP4.
People who already have the footage and need it reformatted: podcast and interview clippers, talking-head creators repurposing a long take, anyone whose horizontal video needs to become a vertical one with captions. It is the only core template that starts from your own camera rather than from text.
It does not find or make the footage for you, and it does not write anything. There is no AI voice on this template; the voice is the one already in your clip. It will not clip a long video down for you either, so upload the cut you want, and it has no free preview because it has nothing to preview until you have uploaded something.
Voice, images, scripts and captions are included on every plan, within fair-use limits. AI credits pay for motion, premium models and extra redos.
Perfect for getting started
Perfect for regular content creators
For professional content creators
Need more mid-month? $9.99 adds 5 videos + 20 AI credits, and never expires.
Custom solutions for teams and agencies
Real results. Real reach. Built with GhostShorts.
Made a Reddit story short in the middle of my chem class. 100K views before I got home.
Honestly thought this was just another overhyped AI tool. But it actually delivers and is super easy to use. I already created 10 shorts in 1 day.
Tried a couple viral formats on the app. A month later... 10 million views. This app is stupid easy to use.
Making edits CapCut felt homeworkish. GhostShorts is easy af, almost feels likes I am on autopilot.
All I do is copy a Reddit post, paste it in, export, done. It's stupid easy.
Used to doom scroll all night. Now I post the same type of videos I used to binge.
I run a few fake text story channels. Jesus I've hit over 5 million likes on just one of my channels alone. It's ridiculous.
GhostShorts saved me countless hours of editing time. Highly recommended!
I cut my content creation time by 80% after switching to GhostShorts. It's the only tool that actually keeps up with how fast I work.
We are here to help you with any questions you may have. If you don't find what you need, please contact us at [email protected]
The same template runs through the GhostShorts API and MCP server. Connect it once and ask your assistant for a video; it comes back as a finished MP4 on the same plan and credits.
claude mcp add --transport http ghostshorts https://api.ghostshorts.com/mcp --header "Authorization: Bearer <your API key>"The split-screen format works on a simple idea. When a video has one thing to look at, any slow second is a reason to scroll. Add a second moving layer, gameplay on the other half, and the eye always has somewhere to go, even while the speaker pauses, thinks or sets up the point.
That is exactly why it suits real footage. A podcast clip, an interview or a talking-head take has natural pauses that would lose a viewer in a full-frame edit. With gameplay running beside it, the pause stops being dead air. Your clip carries the message, the background carries the eye, and the captions carry everyone watching with the sound off.
Gameplay works because of movement. Subway Surfers, Minecraft parkour and racing footage are constant, fluid motion with bright colors and no dead frames. People track movement without deciding to, so the background half keeps them in place while your clip does the talking.
The format is flexible. The best creators pair the clip they want people to hear with a background that fits its mood and pace. These combinations come up again and again.
| Your clip | Background | Good for |
|---|---|---|
| A podcast or interview moment | Subway Surfers | Hot takes, guest stories, the one line everyone quotes |
| A talking-head explainer | Minecraft parkour | Tips, how-tos, longer explanations |
| A storytime take to camera | Minecraft parkour | Confessions, “you will not believe what happened” stories |
| Commentary or an opinion | GTA V or Mario Kart driving | Rants, motivation, strong opinions |
| Calm advice or a tutorial | Satisfying clips | Education, audiences that are not into gaming |
The background does not need to relate to your clip. The contrast between a serious conversation and casual gameplay is part of why the format works: people follow the words through the audio and captions while their eyes rest on the movement. If what you have is a script rather than footage, a narrated template is the better fit: Reddit Story and Roblox Rants read your text with an AI voice over gameplay.
Choose Top/Bottom, the classic short-form split, or Left/Right. Your main clip is the half that carries the audio.
Each clip has its own position controls, so you can slide a horizontal interview until the speaker’s face sits in the middle of their half instead of being cropped off the edge.
Turn on Blur background and a blurred copy of your own clip fills the frame behind it, with your content sharp on top. It is the clean way to turn a horizontal video into a vertical one when gameplay would not suit the audience.
Press Generate AI captions and the speech in your clip is transcribed with every word timed to when it is said. Fix any word in the caption timeline, pick a caption style, and place the captions at the top, the middle or the bottom.
The background matters more than most creators expect. The right one keeps people watching; the wrong one competes with your clip. The library covers Minecraft, Subway Surfers, GTA V, Mario Kart, Roblox, satisfying clips and more, and you can upload your own.
Bright, saturated colors that pop on a phone screen, a character that never stops running, and nothing violent or disturbing on screen. It works as ambient movement rather than a focal point, so it never fights the captions for attention.
The blocky look is recognized everywhere, and the jump-to-jump rhythm pairs well with speech. It feels a little calmer than Subway Surfers, which suits longer, more thoughtful clips.
For audiences that are not into games, satisfying footage gives the same passive pull without the gaming association. It suits education, advice and news-style clips.
If your niche has its own visual language, such as cooking, cleaning or your own gameplay, upload that as the background instead. It can feel more native to a lifestyle or home audience than a mobile game.
The general rule: fast clips with fast gameplay, calm clips with calm footage. A punchy hot take sits well over Subway Surfers or driving; a slow, reflective conversation sits better over parkour or satisfying clips. If people watch the background more than they listen to your clip, the mismatch is too big.
A split-screen video has two sets of rights in it, and they are worth thinking about separately. This is general information, not legal advice.
This is the half that matters most. You need the right to use the footage you upload. A clip from someone else’s podcast is still their work, and putting gameplay under it does not change that. Systems such as YouTube’s Content ID match audio and video against reference files, so someone else’s audio can be matched whatever sits in the other half. Your own recordings, or clips you have permission to use, avoid the problem.
Game publishers set their own rules for videos of their games, and they differ. Mojang’s published usage guidelines, for example, allow videos of Minecraft gameplay, including monetized ones. Check the policy of any game you use, especially footage you upload yourself.
In the United States, fair use can protect a use that transforms the original, and adding your own commentary and a second layer helps that case. But fair use is decided case by case, and it is a defense rather than a permission. The safest split-screen video is one where you own or have permission for everything in it.
You can build a split-screen video yourself in CapCut, Premiere Pro or DaVinci Resolve, and many creators start that way. Here is how the two approaches compare.
| Manual (CapCut, Premiere, DaVinci) | GhostShorts | |
|---|---|---|
| Setup | Import assets, build a timeline, stack two layers | Upload your clip and pick a background |
| Gameplay | Record it yourself or hunt for usable clips | Pick from the library, or upload your own |
| Layout and framing | Resize and position each layer by hand | Choose Top/Bottom or Left/Right, then slide each clip into place |
| Captions | Type and time every word, or run a separate caption tool | Transcribed from your audio, timed per word, editable |
| Render | Export settings to get right every time | 1080x1920 MP4 at 30 fps, in about a minute |
| Cost | CapCut has a basic version without a subscription; Premiere and DaVinci Studio are paid | Starter is $19.99 a month for 30 videos |
| Consistency | Varies with skill, attention and fatigue | Same layout and caption style on every video |
The difference shows up over a week. Posting one split-screen video a day by hand means a daily editing session; with GhostShorts the work is choosing the moment worth clipping. See the pricing page for current plans.
Making a split-screen video is simple. Making one people finish comes down to a few details.
The most overlooked factor. A clip that builds slowly wants calm footage; a fast, punchy clip wants fast gameplay. When both halves share an energy level, people stay longer without knowing why.
Plenty of people watch short-form video with the sound off, so for them the captions are the video. Pick a bold, high-contrast style and a position where the words do not cover a face.
The first second decides whether someone watches or scrolls. Trim your clip before you upload so it opens on the best sentence, not on “so, yeah, anyway.” No intros, no logos, no slow build. For chat formats, the first message plays the same role, as on the fake text message video generator.
Do not assume Subway Surfers always wins. Post the same clip over different backgrounds and compare retention after a couple of days. Swapping the background in GhostShorts is a quick change, not a re-edit.
The layout works because neither half ever gets boring. Past about a minute, that stops being true. Cut to the moment you actually want people to hear.
Give the last few seconds a job: “follow for part 2,” “which one would you pick?” or “comment if you agree.” Comments and follows tell the algorithm the video worked. Do not just let it stop.
Pick a background, generate the captions, render. About a minute per video.
Start Creating Today