HeyGen AI digital human video complete tutorial, 2026 marketing short video in 7 steps

HeyGen is the hottest AI digital human video tool of 2026, with cumulative users surpassing 50 million, used for marketing short videos, corporate training, and social media content production. Unlike video tools such as Runway or Sora that generate footage from scratch, HeyGen's core is to automatically turn a piece of text or audio into an AI digital human explainer video featuring a lifelike presenter, synced lip movements, and facial expressions. This article walks you from sign-up to producing your first finished video in 7 steps.

Beginners most often ask three things. First, does HeyGen require payment to use? Second, how is its Chinese-language support? Third, can the generated videos be used commercially? We address all three below.

HeyGen's Product Positioning and Pricing

HeyGen's official site is heygen.com. Founded in 2020 with headquarters in Singapore, it reached a valuation of $500 million in 2024. Its core capability is generating digital human videos from text.

The free plan offers a monthly quota of 3 minutes of video at 720p with a HeyGen watermark. It is suited to trying things out first.

The Creator plan, at $24 a month or $240 a year, offers 15 minutes of video at 1080p, no watermark, and support for more than 70 digital human avatars. It suits individual creators.

The Team plan, at $69 per person per month, offers 30 minutes, 4K, support for cloning custom digital human avatars, a brand library, and team collaboration.

The Enterprise plan requires contacting sales and offers unlimited duration, API access, and SSO login, suited to large enterprises.

Users in mainland China pay with an overseas credit card or a OneKey-type virtual card. Monthly charges are converted at the bank's daily exchange rate.

Step 1: Register an Account

Open heygen.com and click Sign Up in the top right. You can register via a Google email, a Microsoft account, or email and password. After signing up, you will be asked to enter a use case such as marketing, training, or content creation; this is for HeyGen to tune its recommended templates and does not affect functionality.

After signing up, you receive a free 3-minute quota, and new users get 50% off when upgrading to Creator in their first month. If you only want to try it out, use the free quota to make 2 to 3 short 30-second videos and test the results.

Accessing HeyGen from a mainland China IP works and is not blocked. However, some enterprise-tier features such as the Avatar 4 HD version need an overseas node to load reliably.

Step 2: Choose a Digital Human Avatar

After logging in, you enter the main Studio interface. Click Avatars in the left menu to browse the digital human library. HeyGen has more than 700 built-in digital humans, categorized by gender, age, style, and scenario.

Business style: in suits, suited to corporate promotion and product explanations. Representative avatars include Andrew, Susan, and Maria.

Casual style: in everyday clothes, suited to short-video product promotion and lifestyle accounts. Representative avatars include Jacky, Linda, and Aaron.

Asian faces: HeyGen added more than 200 Asian digital humans in 2024, with Chinese, Japanese, and Korean appearances. For content made in China, choosing Asian faces raises audience acceptance by 30%.

Each digital human comes in three shots: half-body, full-body, and close-up. The half-body shot is most common and supports gestures. The full-body shot suits product demonstrations. The close-up suits interviews.

Preview a digital human's sample speech video, listen to the texture of its English and Chinese pronunciation, and see which one best fits your brand positioning. Once chosen, click Use to enter the editor.

Step 3: Write a Script or Upload Audio

In the editor you see a central canvas, with the script input area in the bottom left. Use one of these methods.

Method 1: Type directly. Type what you want to say into the input box. About 500 Chinese characters correspond to roughly 3 minutes of video. Both Chinese and English are supported. HeyGen automatically segments the text by sentence and generates lip movements.

Method 2: Upload audio. If you have already recorded your own voice, upload an MP3/WAV file. HeyGen automatically syncs the digital human's lips. This method suits keeping the host's own voice while using a digital human's visuals.

Method 3: Voice cloning. Supported on Creator and above. Record 1 minute of your own voice and upload it; HeyGen trains your voiceprint, and subsequent scripts are read aloud in your cloned voice. Voiceprint training completes within 24 hours.

For users in China writing Chinese scripts, HeyGen's Chinese TTS uses the ElevenLabs multilingual version, and by 2026 its Chinese pronunciation is close to a real person in naturalness. However, colloquial filler sounds will be read aloud, so for formal videos it is best to remove them.

Step 4: Set Up the Background and Elements

Now that you have a digital human and a scripted voiceover, it is time to add packaging.

Background: click Background in the top right to switch. HeyGen has more than 200 built-in backgrounds—office, café, outdoor, solid color, and green screen. You can also upload a custom background image. Background images are best at 1920x1080 landscape or 1080x1920 portrait.

Captions: click Captions on the left to add them. Captions are auto-generated in Chinese or English, with adjustable font, color, and position. We recommend a font size of 36 to 48, white text with a black outline, placed in the bottom 15% of the frame.

Logo: Team plan and above support one-click placement from the brand library. On the Creator plan, manually upload a PNG and place it in the top right.

B-roll insertion: when you need to show a product image, data chart, or screenshot mid-video, click the + button to upload an image or video clip and drag it to the corresponding second. HeyGen automatically inserts it center-frame or as a picture-in-picture.

Transitions: each segment defaults to a fade. You can switch to fade, zoom, slide, or glitch styles.

Step 5: Preview and Adjust

Click Preview in the top right to see the result. The preview is 720p; only the full finished render produces 1080p or 4K.

Fixing inaccurate lip-sync: if a line's lip movements do not match, change the punctuation in the script (for example, turn a period into a comma), or insert [pause 0.5] between sentences to force a break.

Fixing mispronunciations: if a Chinese name or place name is read incorrectly, substitute pinyin in the script. For example, if "Mbappé" is read wrong, change it to [mu ba pei]. HeyGen recognizes pinyin and forces the corresponding pronunciation.

Speed adjustment: the slider above the script controls speed from 0.5 to 2.0x. For Chinese we recommend 1.0x, and for English 1.1x, which sounds more natural.

Emotion tags: HeyGen added emotion tags in 2026. Add [happy], [serious], or [excited] at the start of a paragraph, and the digital human adjusts its expressions and movements.

Step 6: Generate the Final Video

When satisfied with the preview, click Generate in the top right. HeyGen composites the script, voice, and visuals into a complete video.

Generation time: a 1-minute video takes about 5 to 8 minutes to render. A 3-minute video takes 15 to 25 minutes. During peak hours you may wait longer.

Queue priority: Creator gets normal priority, Team gets medium priority, and Enterprise gets the highest priority. Free users get the lowest priority.

Background processing: you can close the web page during generation, and HeyGen will email you when it is done. You can also keep the page open to watch the progress bar.

Download format: MP4 with H.264 encoding, defaulting to 1080p 24fps. You can download 4K 60fps on Team plan and above.

File size: a 1-minute 1080p video is about 30 to 80 MB, suitable for uploading directly to Douyin, Bilibili, or YouTube.

Step 7: Publish to Platforms

After downloading the generated video locally, choose your publishing channel.

Domestic platforms: Douyin, WeChat Channels, Bilibili, and Xiaohongshu. All four support direct MP4 upload. Note that Douyin disallows excessive watermarks; a free-plan video with the HeyGen watermark may be flagged as reposted content, so upgrading to Creator to remove the watermark is recommended.

Overseas platforms: YouTube, TikTok, and Instagram Reels. HeyGen has a built-in one-click publish to YouTube feature, supported on Creator and above.

Vertical editing: Douyin and Xiaohongshu use the 9:16 vertical format. Choose the Vertical template when editing in HeyGen, and the export will be 1080x1920. A landscape video needs a second round of editing to change its dimensions after export.

Cover image: HeyGen automatically takes a screenshot of the first second as the cover. You can also upload a custom cover PNG.

Analytics tracking: after publishing, check plays, likes, and retention data in the platform's backend. HeyGen itself does not track data on external platforms.

What Scenarios Is HeyGen Suited For

Marketing short videos: e-commerce merchants use digital humans to explain product selling points. A 3-minute video can be made in 5 minutes, at far lower cost than hiring real people to film.

Corporate training: new-employee onboarding, compliance training, and product training. A single production can be reused, with no need to specially hire instructors.

Social media content production: content creators use digital humans for daily videos, with one person handling content, editing, and publishing end to end.

Foreign-trade customer development: send personalized welcome videos to overseas clients. HeyGen supports name variables to batch-generate 1,000 videos each addressed to a different recipient.

Educational courses: online education platforms use digital humans to explain courses, reducing reliance on real on-camera instructors.

Game NPC voiceover: game developers use digital humans to quickly generate dialogue videos for testing storylines.

HeyGen's Limitations

Complex movements are not possible. Digital humans can only make basic gestures; actions such as running, jumping, and climbing are not supported. For action videos, you still need real people.

Long videos are not cost-effective. Videos over 10 minutes take long to generate and cost a lot, and viewers easily spot the AI traces and lose interest. HeyGen suits 1-to-5-minute short videos.

Chinese colloquialism is weak. HeyGen's Chinese is already very good, but filler words sound stiff when read aloud. For formal videos it is best to change colloquial words to written language.

Pricing is on the high side. At $24 a month, Creator is pricey within China. If you make no more than 5 videos a month, pay-as-you-go options like D-ID or Synthesia may be better.

Copyright risk. HeyGen's built-in digital humans all carry commercial licenses and can be used commercially with confidence. But when cloning a custom digital human, the uploaded real person's portrait must be authorized by that person; otherwise it infringes their likeness rights.

Network dependence. HeyGen is a cloud tool and does not support offline generation. If your connection is unstable, both uploading material and downloading videos are slow.

Frequently Asked Questions

Can HeyGen's free plan be used commercially?

No. Videos generated on the free plan carry a HeyGen watermark, and the terms of service state that free users do not hold commercial licenses. If you use free-plan videos for marketing, sales, or corporate promotion and HeyGen detects it, you may be warned or even banned. Commercial use requires upgrading to Creator at $24 a month or above. Creator includes a commercial license, but only for your own business; you cannot resell the videos themselves. The Team plan allows you to produce and deliver videos for clients.

Does the Chinese version of HeyGen sync the lips accurately?

About 90% accurate. HeyGen upgraded to Avatar 4 in late 2025, with Chinese lip-sync specially optimized. Ordinary declarative sentences match very naturally. But a few scenarios are prone to problems. First, mixing English into Chinese (for example, "use ChatGPT to write code") can make the lip-sync inaccurate on the English part. Second, number reading—whether 1234 is read as "one thousand two hundred thirty-four" or "one-two-three-four"; HeyGen reads it as an integer by default, so to change it write it as digits read individually in the script. Third, special symbols such as percent and plus signs, which HeyGen reads aloud as "percent" or "plus," sometimes not as expected, so it is best to change them to Chinese words.

Is it safe to upload my own face to make a digital human?

HeyGen's custom avatar feature requires users to sign a likeness authorization consent form, confirming the portrait is yourself or someone you have authorization for. After uploading, the avatar data is stored on HeyGen's AWS servers, which hold SOC 2 Type II certification. In theory the data will not leak, but cautious users can wait for HeyGen's local-processing version to launch. Additionally, you can delete it at any time after uploading; deletion requests clear all copies within 7 days. But note that making a digital human from someone else's face without authorization infringes their likeness rights, and a lawsuit can be costly.

Is access to HeyGen stable within mainland China?

It is accessible, but some features require a VPN. HeyGen's domain heygen.com opens fine in China, and basic registration, login, browsing templates, and writing scripts are all problem-free. But the video-generation stage calls ElevenLabs TTS, AWS video rendering, and CDN downloads, which often stall or time out. We recommend connecting to an overseas node such as Hong Kong, Singapore, or Japan. Cloudflare WARP can also solve part of the problem. For long-term use, buying a stable overseas VPS as a proxy is recommended.

Which is better, HeyGen or Synthesia?

Each has its focus. Synthesia, founded in 2017, specializes in the corporate training market with more than 230 built-in ultra-realistic digital humans, suited to internal corporate training and compliance videos. HeyGen has a larger digital human library of more than 700, updates faster, and suits marketing short videos and social media. On price, HeyGen Creator at $24 is cheaper, while Synthesia starts at $29 a month. HeyGen's Chinese support is slightly stronger. For large-scale corporate training, choose Synthesia; for marketing and content creation, choose HeyGen. The two tools serve different user groups with little conflict, and large companies often buy both.