How to make AI videos as a beginner: pick a starting point, choose an avatar, add a script, and publish in 6 steps. Includes free plan limits. Start free.
TL;DR: The fastest path for a beginner is HeyGen's Video Agent, because one typed prompt produces a finished video with a script, an AI presenter, voiceover, captions, and motion graphics, with no editing skills or camera. The Free plan lets you make up to 3 short videos a month to learn on.
- Synthesia is a good fit if your first videos are corporate training built on slides.
- Canva works for simple text-and-stock-footage social clips when you already design there.
- CapCut is the better pick when you're editing footage you filmed yourself.
To make an AI video, type what you want into an AI video generator, pick a presenter and voice, add your script, and render. A first video takes about 15 minutes.
The hard part for beginners isn't the software. It's knowing which of five starting buttons to press, why the preview looks wrong, and how far a free account goes. This guide answers all three in plain steps.
You'll learn the six-step workflow, what the Free plan can and can't make, and the mistakes that waste a first month. HeyGen is rated 4.8/5 on G2 from 1,901 reviews, and G2 ranks it #1 for most realistic avatars.
How to Make AI Videos with HeyGen
1. Sign up and pick a starting point
Create a free account, no credit card required. On the dashboard you'll see several ways to start: Video Agent (type a prompt), templates (fill in a ready layout), a blank project in AI Studio, and import options for PowerPoint or PDF.
Beginners should start with Video Agent for a first video, because it makes every decision you haven't learned yet. Move to templates or AI Studio once you want control over each scene. Signing up takes about 2 minutes.
2. Choose a presenter
Pick a stock avatar, upload a photo to make a photo avatar, or record yourself to build a Digital Twin. Free accounts get 500+ stock video avatars, up to 3 photo avatars, and one custom video avatar slot.
Stock avatars are ready instantly. A Digital Twin needs about 30 seconds of footage plus a consent clip and is usually ready in 10 to 20 minutes, and an AI photo avatar needs one front-facing image.
3. Add your script or prompt
Type or paste your script into the script box, or upload your own voice recording. In Video Agent, describe the video instead, and start with the length: "Make a 45-second video explaining…"
Aim for about 150 words per minute of finished video. A one-minute free-plan video holds roughly 150 words, so write short sentences that sound natural read aloud. Allow 5 to 10 minutes for a first script.
4. Pick a voice and build the scenes
Choose a voice that matches your audience's language and accent. Voice cloning lets your presenter speak in your own voice, and the Free plan includes one clone.
In AI Studio, add text, images, logos, music, and B-roll scene by scene, or apply a Brand Kit for your colors and fonts. For a first video, two or three scenes are plenty. Building them takes about 5 minutes.
5. Preview, then generate
Click preview to check timing, text placement, and scene order. Expect the avatar's mouth to look blurry in the preview: lip sync only renders when you submit the video, so it isn't a bug.
When everything else looks right, click Generate or Submit. Short videos usually render in minutes. While you wait, the video processing guide explains what to do if a render takes longer than expected.
6. Download, share, or translate
Download the MP4 or share a link. Free exports include a watermark, paid plans remove it and export at up to 1080p on Creator or 4K on Pro, and every plan offers vertical, square, and widescreen formats.
Want the same video in Spanish or Japanese? Paid plans translate it into 177+ languages and dialects with lip sync. The Free plan covers 30+ languages for testing.
What to Have Ready Before You Start
- One clear topic. A single question your video answers. "How to reset your password" beats "Everything about our app."
- A short script or outline. 150 words for a one-minute video. Bullet points are fine for Video Agent.
- A destination. Know where the video will live, because TikTok needs vertical 9:16 and YouTube or training needs widescreen 16:9.
- Brand basics. A logo file and one or two brand colors. Good enough means a PNG and a hex code.
What a Free Account Can Make
Most beginner guides say "start free" without saying what free includes. Here is what the current pricing page lists, so your first month has no surprises.
On paid plans, credits are spent per minute of finished video, and the rate depends on the avatar model. HeyGen's avatar page lists Avatar IV at 16 to 31 credits per minute and Avatar V at 48, so Creator's 600 credits cover about 12 minutes of Avatar V.
If you're on an older plan, the rules differ. The post Is Avatar III still unlimited explains how legacy and credit-based plans compare.
Quick Video vs. Polished Video
Depending on your goal, there are two ways to make the same first video.
The 10-minute version
Open Video Agent and type: "Make a 45-second vertical video introducing my bakery's new sourdough to Instagram followers, friendly tone." Let the agent pick the presenter, voice, and visuals. You'll get a shareable video and a feel for what the tool does.
The polished version
Write your own 150-word script, build a Digital Twin so it's your face, clone your voice on a paid plan, apply your Brand Kit, and assemble three scenes in AI Studio. Budget about an hour the first time. Use this route for videos customers or colleagues will see more than once.
Common Mistakes to Avoid
Writing a script that's too long for the plan
The Free plan stops at 1 minute per video. A 400-word script won't fit, so cut it to about 150 words or split it into three short videos.
Panicking at the blurry mouth in preview
New users often think lip sync is broken because the preview mouth looks soft. It sharpens in the final render, so generate a short test before changing anything.
Choosing a presenter for looks, not audience
A presenter in a home kitchen won't suit a compliance video, and a formal studio presenter can feel stiff in a social ad. Match the setting and outfit to where viewers will watch.
Spending credits on long drafts
Test with a 20 to 30 second excerpt before rendering a full script on a premium model. Fixing a pronunciation in a short test costs a fraction of re-rendering the whole video.
Other Ways to Make AI Videos
Synthesia
Synthesia is an avatar platform built for corporate training.
- Pros: a slide-style editor familiar to PowerPoint users; consistent presenters for training; one-click translation; a long enterprise track record.
- Cons: the Free plan includes only 9 avatars; Starter caps output at 10 minutes of video a month, which limits experimentation for beginners.
Canva
Canva's video editor combines templates, stock footage, text, and music, with AI features layered on top.
- Pros: a familiar interface for most marketers; huge template library; strong for social graphics and short clips; easy team sharing.
- Cons: no realistic speaking presenter with lip sync, so teaching and sales videos feel impersonal; stock footage appears in many other brands' content.
CapCut
CapCut is a video editor popular with short-form creators, with AI captions and effects.
- Pros: fast editing for TikTok and Reels; auto-captions; trending effects and templates; free to start.
- Cons: it edits footage you already have, so you still need to film; no AI presenter that speaks your script.
Frequently Asked Questions
Can I make AI videos without showing my face?
Yes. Use a stock avatar or a photo avatar of an illustrated character, and your face never appears, which is how many faceless channels and training teams work. If you want your face without filming each time, a Digital Twin needs one short recording.
How long does it take to make a first AI video?
About 15 minutes with Video Agent, including writing the prompt and waiting for a short render. A polished video with your own Digital Twin and Brand Kit takes about an hour the first time and much less once your setup is saved.
Can I use AI videos for my business?
Yes, on eligible paid plans. Videos can be used for ads, marketing, social media, and client work, and stock avatars are cleared for commercial use. The Free plan's watermark makes it a testing tier rather than a publishing tier, so upgrade before anything goes to customers.
Can two avatars talk to each other in one video?
Yes. You can build a conversation with multiple scenes, Avatar Shots, or the API. The guide on how to make two avatars talk walks through each method, and the Video Podcast app builds a two-host show from a single script or document.
What's the difference between Video Agent and AI Studio?
Video Agent builds the whole video from a prompt, making decisions on scenes, visuals, and pacing for you. AI Studio is the editor where you place every element yourself. Most people start a video in Video Agent and fine-tune it in AI Studio.
Conclusion
To make AI videos as a beginner, start in Video Agent with a short prompt, pick a presenter, keep the script near 150 words per minute, and test before rendering long drafts. The Free plan covers your first three videos, and Creator starts at $29 per month.
Greetings! My name is Ayesha Shaheryar. My words have helped millions over the past two years. As a HeyGen expert and a writer, I am here to introduce tips and tricks to edit your next video in no time.
