How to make AI videos as a beginner
A beginner can make an AI video in VideoSlash from a topic: a script, a voiced faceless short with captions, or a custom clip from a photo. Then download it or post it to TikTok.
To make an AI video as a beginner, start with the finished video you want, not with a model name. In VideoSlash that is usually a short with a voice and captions: you give it a topic, it drafts the script, then it builds the video. If you already have a photo or a specific shot in mind, Creative Studio turns a prompt or a still into a clip. If the pieces need to be cut together, the Pro Editor and its agent do that on a timeline.
You do not need filming experience. You do need one clear idea, and you need to watch the result once before you post it.
Which VideoSlash tool should a beginner open first?
| What you want | Open this | What you leave with |
|---|---|---|
A narrated short, no camera | Pro Studio, faceless video | Voice, visuals, and captions in one file |
A text-message story | Fake text video | The conversation playing on screen |
A narrated post-style story | Reddit-style video | A story video with voice and captions |
A clip from a sentence or a photo | Creative Studio | An image, an edited image, or an image-to-video clip |
Several clips cut into one export | Pro Editor | A timeline, with an agent that can trim, caption, and export |
The same jobs from Claude, Cursor, or Codex | MCP or the public API | Scripts and Creative Studio generations from the chat you already use |
Premium is €39 a month or €297 a year, with 100 credits refreshed every month. Script generation on that plan is unlimited. Details are on the pricing page.
What “making an AI video” actually is
There are two different jobs, and beginners mix them up.
The first job is a finished short. Someone speaks, pictures change with the lines, captions sit on the screen, and the file is vertical so TikTok, Reels, and Shorts can use it. VideoSlash builds that in Pro Studio from a topic or a script you paste in. Fake text and Reddit-style videos are the same idea with a different layout on screen.
The second job is a generated shot. You describe a scene, or you upload a picture and ask for motion. That is Creative Studio: text-to-image, image-to-image, image-to-video, and edits. The clip is a piece. It does not yet have a full voiceover script and a publish step unless you add those.
Most people who want “an AI video for TikTok” want the first job. Start there. Use Creative Studio when one line in the video needs a picture the template did not make.
What you need before you generate
Three things.
A single idea. One mistake, one story, one product, one question. “Personal finance” is a niche. “I paid $40 a month for an app I opened twice” is a video.
A first line someone would not scroll past. A number, a question, or a conflict. “Welcome to my channel” is not a first line.
A source image, only if the shot has to match a real object. A product photo, a room, a character still. Sharp and uncluttered works better than a dark, busy frame. You upload that in Creative Studio for image-to-image or image-to-video.
A prompt for a custom shot needs four concrete details: who or what is in the frame, what they do, where they are, and how the camera moves. “A woman walking” gives the model nothing to hold onto. “A woman in a red coat walks through a rainy street at night, close-up, neon on the wet pavement” gives it a shot.
Skip words like beautiful, amazing, and epic. They do not change the picture.
Make your first faceless video
- Open the script generator. Type the topic in one sentence, or paste a script you already like. Premium drafts the voiceover for you. Read the first line out loud and replace it if you would skip it.
- Open Pro Studio and generate the faceless video from that script. The render includes the voice, the visuals, and the captions. The plan includes 50+ voices, 100+ languages, and 15+ caption styles, so you can change the voice or the caption look without rewriting the script.
- Watch the whole file with the sound off. If the captions still tell the story, it will work on a phone in a loud room. If a picture stays up after the sentence has moved on, that is the spot to fix.
- Download the file. In the dashboard you can also publish or schedule it on TikTok. Instagram Reels and YouTube Shorts take the same download. You upload those yourself.
That is a complete AI video. No camera, no timeline, unless you want one.
Fake text starts from the conversation instead of a narration. You set the messages, names, and timing. Reddit-style starts from the story you want read aloud. The download and the TikTok post work the same way.
If you already like how another YouTube video is structured, Pro Studio can clone a script from that URL. You get a new script in a tone and length you choose. You do not get their video, and VideoSlash does not post it to YouTube for you.
Make a custom clip in Creative Studio
Use this when the faceless template is not the shot you want.
- Pick image or video, and pick a model. The catalog is in Creative Studio. You do not need a separate account for each model.
- Write the prompt with subject, action, setting, and camera. For a product or a face that must stay put, upload a still and use image-to-video or image-to-image instead of text alone.
- Generate. The result stays in your Creative Studio library. A failed generation shows a failed tile and the reason, and the credits for that run are refunded. Some prompts are rejected because the provider flags them. Change the inputs and run it again.
- If that clip belongs inside a longer video, import it into a Pro Editor project. The editor can also generate from a sidebar without leaving the timeline. Those generations stay in Creative Studio until you add them to the project.
Change one thing between tries: the prompt, the source image, or the model. Changing all three means you will not know which one fixed the clip.
Cut it together in the Pro Editor
The Pro Editor is a multi-track timeline, separate from the faceless renderer. Open it when you have more than one piece: a Pro Studio video plus a Creative Studio shot, or a caption you want to fix by hand.
You can edit on the timeline yourself. You can also use the editor agent in the project chat. It can look at the timeline, trim, split, add titles and captions, generate media onto the timeline, and export. Director mode plans the cut and can switch into design, audio, editing, or story for a step. Captions in the editor can be transcribed for one clip or for the whole timeline.
Export when the cut is done. Then publish or schedule TikTok from the dashboard, and upload the file everywhere else.
Let an agent do the repetitive parts
The MCP server and the public API use the same account. Claude connectors sign in. Claude Code, Cursor, Codex, and OpenCode use an API key from Account, Developers. The key starts with vsl_.
You can ask for the credit balance, a faceless script, or a Creative Studio image or video. A first prompt that works: “Check my credits. Write a short faceless script about paying for an app I never open. Then generate a 5-second clip of a phone on a kitchen table.”
Script generation on Premium does not spend the 100 monthly credits. Image and video generation does. If both balances are empty, the tools tell you instead of pretending the job ran.
The in-app editor agent is separate from MCP. It runs inside the Pro Editor and spends editor-agent credits, or it can use an Anthropic key you connect. MCP is for the chat you already work in. The editor agent is for the timeline that is open.
A beginner’s first week
Stay on one format. Three faceless explainers, or three fake texts, or three Reddit-style stories. Mixing them makes three channels.
- Video one: a specific loss or mistake in the first line.
- Video two: a list of three, with the useful item last.
- Video three: a story that turns in the last sentence, so a replay makes sense.
Each one goes topic, script, generate, mute-watch, download, TikTok. Use Creative Studio only for the shot the template missed. Use the Pro Editor only when you are joining pieces.
After posting, check the file, not the view count. The spoken first line should match the post caption. The captions should work with the sound off. The picture should change when the sentence changes. The last line should pay off the opening.
What beginners get wrong
| Mistake | What you get | What to do instead |
|---|---|---|
Starting from a model name | A clip and no script, voice, or captions | Start in Pro Studio if you need a short you can post |
A vague topic | A generic script | One concrete fact in the first line |
“Beautiful” and “cinematic” in the prompt | The model guesses | Subject, action, setting, camera |
Text-to-video for a specific product | A different object every time | Upload the photo and use image-to-video |
Posting without a mute watch | Captions that do not match the voice | Watch once with the sound off |
Changing prompt, image, and model together | No idea what helped | One change, then generate again |
What it costs
Premium is €39 a month or €297 a year. Both include 100 credits a month, unlimited script generation, Pro Studio, fake text, Reddit-style videos, Creative Studio, the Pro Editor, and API and MCP access. There is no watermark, and commercial use is included.
A script does not spend those 100 credits. Generating images and video does. A failed provider run refunds the credits for that run. You can see the balance on the dashboard.
You own the videos you create. The FAQ allows them on YouTube, TikTok, and Instagram, for client work, and in ads, within the terms of service. TikTok can be published or scheduled from VideoSlash. YouTube and Instagram are a manual upload of the same file. Those platforms still decide who gets into their payout programs.
Questions beginners ask
Do I need technical skills?
No. The dashboard is a form: a topic or a script, then generate. Creative Studio is a prompt, an optional image, and a model. MCP is a sentence in Claude, Cursor, Codex, or OpenCode after you connect the account.
What is the difference between a faceless video and image-to-video?
A faceless video is a full short: script, voice, pictures, captions. Image-to-video animates a photo you upload in Creative Studio. Use image-to-video when the object or the face has to match a picture you already have.
How long should a prompt be?
For a script, one specific topic and a strong first line. For a Creative Studio shot, one or two sentences that name the subject, the action, the place, and the camera. A short specific prompt beats a long vague one.
Why does the clip not match what I wrote?
The model fills in anything you left blank. Name the action, the place, and the framing. Drop empty adjectives. If the subject is a real product, upload the photo instead of describing it.
Can I use a real person’s photo?
You can upload a still for image-to-image or image-to-video. Providers reject some inputs and some outputs. When that happens, the tile fails and the credits for that run come back. Follow VideoSlash’s terms and the platform rules before you post anyone’s face.
Where do I post the video?
Download it for Instagram and YouTube. Connect TikTok in the dashboard to publish or schedule there. The file you want for those feeds is vertical.
Can Claude or Cursor make the video for me?
Yes, after you connect MCP. They can check credits, write a script, and run a Creative Studio generation. The editor agent, inside a Pro Editor project, can cut and export the timeline.
Start with one topic and a faceless video on videoslash.com. Add a Creative Studio shot only when the template picture is wrong. Post that one file before you open a second project.