WTF AGENTS
Guide 27 / 37

WTF is HeyGen

A presenter, a script, 175 languages. No camera.

A presenter who never needs a camera, a script that becomes a finished video from one prompt, and dubbing into 175 languages with the lips to match. What HeyGen does, what the credits really cost, and when you should still film yourself.

The one-liner

HeyGen makes videos of people who were never filmed. Type a script, pick a presenter — a stock avatar or a digital twin of yourself made from two minutes of webcam footage — and out comes a talking-head video with natural gestures, lip-sync and, if you want, the same video in 175 other languages.

It is the presenter seat in a solo founder's stack, and for marketing and training teams it replaced the studio booking. In 2026 it added the agent: one prompt, and it writes, voices, shoots and edits the whole thing.

By the numbers — September 2026

$200M+annualised revenue, announced by the company in June 2026 — doubled in eight months
100,000+businesses using it, per the company; G2's fastest-growing product of 2025
175+languages and dialects for translation with lip-sync
2 minof webcam footage to create a digital twin of yourself
$29 / $49 / $149per month for Creator, Pro and Business; free tier of three watermarked videos
31–48 creditsper minute of Avatar IV or V video, against 600 credits on Creator — twelve to nineteen finished minutes a month
$500Mlast disclosed valuation, at the 2024 Series A led by Benchmark

What it actually does

Avatars — Five hundred-plus stock presenters, or your own. Avatar IV and the newer Avatar V read the emotional register of the audio and generate matching expressions, head turns and hand gestures; the stiff mannequin of two years ago is gone.

Digital twins — A custom avatar from two minutes of footage; a photo avatar from a single still. Your face, delivering scripts you never read aloud.

Video Agent — The 2026 headline. One text prompt becomes a complete video: script, visuals, presenter, voice-over, edit. Under four minutes to a finished, postable piece. Speed and precision modes for quick clips versus high-stakes content.

Translation — Upload a video; get it back in another language, in a matching voice, with the mouth movements re-synced. 175-plus languages, from about thirty source languages.

Interactive avatars — A presenter that answers questions live, for sales pages, kiosks and support. A separate LiveAvatar track, billed by the real-time minute, with extra slots at $49 a month.

Brand kits, templates, SCORM — The corporate plumbing: consistent styling, reusable layouts, exports that drop into learning systems.

API and MCP — A standalone pay-as-you-go pool: buy the credits you want, no subscription, and they last twelve months. Studio video, translation and LiveAvatar, so other agents can make HeyGen videos on your behalf.

What it really costs — the credit maths

Every plan is a credit budget, and the credit rates are the part the pricing page does not put in the headline.

PlanPriceCredits and what they buy
Free$03 videos/month, watermarked, one minute max
Creator$29/mo ($24 annual)600 credits, 1080p, videos to 30 min, voice cloning, unlimited photo avatars
ProFrom $49/mo1,000 credits, rising in tiers; 4K export, advanced models
Business$149/mo + $20/seat1,500 credits, collaboration, brand controls, SCORM
EnterpriseCustomStudio-grade avatars, seats, security review, support
What you makeCredits per minuteOn Creator (600)
Avatar III, video look4~150 minutes
Avatar III, photo look7~85 minutes
Avatar IV, photo look16~37 minutes
Avatar IV, video look31~19 minutes
Avatar V, either look48~12 minutes
Video Agent, standard40~15 minutes
Video Agent, Seedance120~5 minutes
Translation with lip-sync6~100 minutes
Translation, no lip-sync4~150 minutes

The plan says $29. Nineteen minutes of Avatar IV is what $29 buys, and twelve if you want Avatar V. A creator making twenty three-minute videos a month needs sixty finished minutes — roughly three times a Creator plan — so that is Pro at a higher tier, and they will still watch the meter.

The API has its own credit pool; web-plan credits do not cross over.

HeyGen vs Higgsfield vs ElevenLabs

HeyGenHiggsfieldElevenLabs
The seatPresenterGenerated videoVoice
Core outputA person talking to cameraClips, ads, motionSpeech, dubbing
One-prompt agentVideo AgentSupercomputerElevenAgents (calls)
Languages175+ with lip-syncDepends on model70+ generation, ~30 dubbing
Entry price$29/moCredits, tiered$6–22/mo
Best forExplainers, sales, training, localisationShort-form, ads, socialAudio of any kind

Most solo founders need one of the three; a marketing team ends up with all of them, often with ElevenLabs voices inside HeyGen videos.

The catches

  • Credits, again. The realistic avatars cost eight to twelve times the old ones per minute. Budget in finished minutes at the quality you will actually publish.
  • The uncanny gap is small, not zero. Avatar IV and V are good enough for corporate content and most marketing; they are not a substitute for a founder on camera when trust is the product.
  • Consent is lighter than at rivals. Security researchers have noted HeyGen's avatar-creation consent checks are less strict than Synthesia's. Cloning anyone but yourself needs their written permission; the platform's leniency is not your defence.
  • Disclosure rules are arriving. Several jurisdictions and platforms now require labelling synthetic presenters in ads and political content. Label where required.
  • Your twin lives there. A digital twin is an asset on HeyGen's servers; extra slots cost $29 a month each. Export finished videos; the avatar does not travel.

Who it is for

Yes — Anyone producing structured, script-driven video at volume: explainers, product demos, course modules, sales outreach, multilingual versions of the same thing. Creator at $29 if it is a few videos a month; Pro if it is a few a week.

Not yet — Anyone whose audience needs to trust a real face, or who is unsure of their monthly volume. Film ten videos yourself first, then decide.

No — Anyone who needs cinematic or generated footage rather than a presenter (Higgsfield, Runway), or only needs the voice (ElevenLabs).

Why it matters

HeyGen is what "AI-run company" looks like at the point of contact with a customer: a presenter that does not exist explaining a product that an agent built, in whatever language the viewer speaks, produced from one prompt in four minutes. The economics of video content changed in 2026, and this is the tool that changed them for the people who make it.

Glossary

HeyGen
AI avatar video platform: digital presenters, one-prompt videos, translation with lip-sync. Los Angeles; $200M+ ARR.
Avatar IV / V
HeyGen's current photorealistic avatar models, with expression and gesture driven by the audio.
Digital twin
A custom avatar of a real person, made from about two minutes of footage.
Video Agent
One prompt to a finished video: script, visuals, presenter, voice, edit.
Video translation
Re-voicing a video in another language with the mouth re-synced. 175+ target languages.
Credits
The usage currency. Rates vary by feature and look: 4 a minute for an Avatar III video look, 31 for Avatar IV, 48 for Avatar V, 40 for the Video Agent.
Interactive avatar
A presenter that responds live to questions. Enterprise, per slot.
SCORM
The export format learning-management systems accept; a Business-plan feature.
Consent
Documented permission from anyone whose likeness you clone. Required regardless of what the platform enforces.

Liked this? Go deeper.

HeyGen is the presenter. Here is the rest of the creative stack.

Get the rest of the series

Every WTF Agents guide is written the same way — plain English, no hype, no jargon. $7 each, or take a bundle.

Go deeper
Get the PDF guide · $7