CtrlK
HeyGen is an artificial intelligence video creation platform that enables users to produce professional videos using AI-generated avatars, voice cloning, and automated video translation. Founded in December 2020 by Joshua Xu and Wayne Liang under the original name Surreal, the company is headquartered in Los Angeles, California. HeyGen allows businesses and individuals to create videos at scale without traditional filming, cameras, or actors. The platform supports over 175 languages and dialects, offers more than 1,100 stock avatars, and serves over 85,000 customers globally. HeyGen reached $100 million in annual recurring revenue (ARR) in October 2025, making it one of the fastest-growing startups in the generative AI space.
HeyGen was co-founded by Joshua Xu (CEO) and Wayne Liang (Chief Innovation Officer) in Shenzhen, China, in December 2020. Both founders attended Tongji University in Shanghai before pursuing master's degrees at Carnegie Mellon University in the United States. After graduating, they relocated to the US West Coast in 2014. Xu spent six years as a software engineer at Snap Inc., where he worked on advertising technology and AI-augmented camera development. Liang worked as a product designer at Smule, a karaoke application.
Xu's experience at Snap exposed him to the challenges of creating high-quality video content, which typically requires on-camera skills, production budgets, and extensive post-production work. Convinced that generative AI could transform visual storytelling, both founders left their jobs in 2020 to launch their own venture under the name Surreal.
In 2021, the company raised a seed round from Sequoia China (now HongShan) and ZhenFund. By March 2021, the platform had accumulated approximately 10 million photo and video orders.
In 2022, Xu and Liang relocated the company to Los Angeles and rebranded it as Movio. The move was motivated by a desire to access advanced semiconductor hardware for training their AI systems and to target major US enterprises such as Salesforce, Amazon, and NVIDIA. The company dissolved its Shenzhen operations in 2023.
In April 2023, the company rebranded once more, this time as HeyGen. This final name change reflected a strategic shift from offering only spokesperson-style videos toward a broader, more comprehensive video generation experience. The name "HeyGen" signals the platform's focus on generative AI technology.
HeyGen launched its web application in September 2022 and maintained a weekly product release schedule from July 2022 onward. The company reached $1 million in ARR within 178 days of its March 2023 commercial push, then scaled to $10 million ARR by October 2023. By June 2024, ARR had grown to over $35 million. The company achieved profitability in Q2 2023 and has remained profitable since.
In 2023, HeyGen expanded its leadership team with several key hires. Dave King, formerly the Chief Marketing Officer at Asana, joined as Chief Business Officer. Rong Yan, previously VP of Engineering at HubSpot, became the Chief Technology Officer. Lavanya Poreddy, with experience at Meta and Match Group, was appointed Head of Trust and Safety.
By September 2025, HeyGen's ARR had reached $95 million, growing to $100 million by October 2025. This milestone was reached just 29 months after the company first crossed $1 million in ARR, representing one of the fastest growth trajectories in enterprise SaaS history.
The core product of HeyGen is its avatar-based video creation platform. Users can select from a library of over 1,100 pre-made stock avatars representing dozens of ethnicities, ages, and speaking styles. Users assign a voice to the avatar (with customizable pitch and speed), input a script, and the platform generates a complete video.
HeyGen offers several types of avatars:
| Avatar Type | Description | Best For |
|---|---|---|
| Stock Avatars | Pre-made avatars from HeyGen's library of 1,100+ options | Quick video creation without personal branding |
| Photo Avatars | Animated avatars created from a single uploaded photo | Turning still images into talking presenters |
| Hyper-Realistic Avatars | Digital twins created from a short user-recorded video | Executive communications, corporate training |
| Avatar IV | Next-generation avatars created from a single photo with expression intelligence | High-fidelity content requiring natural emotion |
Released in April 2025, Avatar IV represents HeyGen's most advanced avatar technology. Unlike earlier avatar models that simply synchronize lip movements to speech, Avatar IV interprets the vocal tone, rhythm, and emotion of the input audio to generate photorealistic facial movements. The system is built on a diffusion-inspired audio-to-expression engine that produces natural head tilts, pauses, subtle cadences, and micro-expressions.
Avatar IV supports angled and profile photos and works with various character styles, including humans, anime characters, and even stylized pets. Users can create up to 500 different "looks" per avatar slot, customizing backgrounds, outfits, camera angles, and stances. The Avatar IV API became available for self-serve use on Pro and Scale tiers, consuming approximately 1 credit per 10 seconds of generated output.
HeyGen's text-based editor allows users to create complete videos by simply typing a script. The platform handles narration, captions, visuals, and animations automatically. Users can control tone, delivery, gestures, and emotion from a single interface, making the process comparable to writing a document rather than producing a video.
The platform provides over 400 video templates spanning categories like marketing, training, sales, and social media. Users can customize these templates or start from scratch.
HeyGen's voice cloning feature allows users to replicate their own voice from a short audio recording. Once cloned, the voice can be used across any video on the platform, maintaining consistent branding and personal identity. The platform also offers more than 300 stock voices across 175+ languages and dialects, with adjustable parameters for pitch, speed, and tone.
Video Translate is one of HeyGen's most prominent features. Users can upload any existing video (or paste a YouTube link), and the platform will automatically transcribe, translate, and dub the video into a target language. The system uses voice cloning to preserve the original speaker's vocal characteristics, including tone, accent, and intonation, while re-synchronizing lip movements to match the translated audio.
The technology works through a multimodal understanding mechanism that simultaneously analyzes audio, transcript, facial expressions, and scene context. This allows it to dynamically adjust the translation style based on visual cues and emotional context. The lip-sync engine handles challenging scenarios including side profiles, obstructed faces, and rapid head movements.
Video Translate supports over 175 languages and dialects, including regional variants such as Argentine Spanish, Mexican Spanish, and Brazilian Portuguese.
HeyGen's Interactive Avatar feature, marketed as LiveAvatar, enables real-time conversational AI experiences. Using computer vision and speech synthesis, LiveAvatar replicates user motions, expressions, and gestures in real time. The feature is designed for customer service, virtual assistants, and interactive educational content.
LiveAvatar is available through an API, allowing businesses to embed real-time avatar conversations directly into websites and applications. The avatars operate 24/7 and combine natural facial expressions, lip sync, and body language to create engaging conversational experiences.
HeyGen provides a developer API that enables programmatic access to its video generation, avatar creation, and translation capabilities. The API supports integration with existing business workflows, CRM systems, and marketing automation platforms.
| API Tier | Monthly Cost | Credits Included | Per-Credit Cost |
|---|---|---|---|
| Free | $0 | 10 credits | N/A |
| Pro | $99 | 100 credits | ~$0.99 |
| Scale | $330 | 660 credits | ~$0.50 |
| Enterprise | Custom | Custom | Negotiated |
HeyGen operates on a subscription-based SaaS model with a credit system for advanced features. As of early 2026, the platform offers the following plans:
| Plan | Monthly Price | Annual Price (per month) | Key Features |
|---|---|---|---|
| Free | $0 | $0 | 1 credit (1 min of video), 1 Instant Avatar, 175+ languages, HeyGen watermark |
| Creator | $29 | $24 | 15+ minutes/month, no watermark, 1080p HD, 3 Instant Avatars, premium voices, voice cloning |
| Business | $149 (+$20/seat) | ~$119 (+$20/seat) | Higher resolution, team workspace, review and collaboration tools, advanced editing |
| Enterprise | Custom | Custom | Unlimited videos, no duration caps, SSO/SCIM, dedicated CSM, custom integrations |
Advanced AI features consume additional generative credits (GenCredits) beyond standard plan allowances:
| Feature | Credit Cost |
|---|---|
| Avatar IV generation | 20 credits/minute |
| AI model training | 20 credits/session |
| Video Translation | 5 credits/minute |
| 4K upscaling | 10 credits/conversion |
The legacy Team plan ($39/seat/month, minimum 2 seats) was deprecated in January 2026. Existing subscribers who maintained uninterrupted payments retained their original terms, but the plan is no longer available to new users.
HeyGen has raised a total of approximately $69 million in venture capital funding across two primary rounds.
The company's initial funding came from a seed round led by Sequoia China (now HongShan) and ZhenFund in 2021. The exact amount of this round has not been publicly disclosed, though it is estimated at approximately $9 million based on the total funding figures.
In March 2024, HeyGen raised $60 million in a Series A round led by Benchmark, which valued the company at $500 million post-money. Benchmark partner Victor Lazarte joined HeyGen's board as part of the investment. Other participants included Thrive Capital, BOND, SV Angel, and Conviction. Angel investors in the round included Neil Mehta, founder and managing partner of Greenoaks Capital, and Dylan Field, CEO of Figma.
The Series A was notable for HeyGen's pivot away from its original Chinese investor base. As the company relocated to the United States and targeted American enterprise customers, it deliberately sought funding from prominent US-based venture capital firms.
HeyGen's revenue trajectory has been among the most rapid in enterprise software:
| Date | ARR Milestone |
|---|---|
| March 2023 | $1 million |
| October 2023 | $10 million |
| June 2024 | $35 million |
| December 2024 | $57.5 million |
| September 2025 | $95 million |
| October 2025 | $100 million |
The company achieved profitability in Q2 2023 and has sustained it since then. As of mid-2025, HeyGen operated with approximately 200 employees, giving it one of the highest revenue-per-employee ratios in the industry.
HeyGen serves over 85,000 customers globally, including more than 500 enterprise accounts. The platform is used across industries for marketing, corporate training, internal communications, customer support, and content localization.
Several high-profile use cases have demonstrated HeyGen's capabilities:
Javier Milei's WEF Speech (January 2024): When Argentine President Javier Milei delivered a speech in Spanish at the World Economic Forum in Davos, a HeyGen-translated version dubbed into English went viral on social media. The translation preserved Milei's voice, accent, and lip movements, accumulating over 75 million views on X (formerly Twitter). The video was shared widely, including by Elon Musk, and brought significant public attention to AI-powered video translation technology.
UNDP Weather Kids Campaign (March 2024): HeyGen partnered with the United Nations Development Programme, the World Meteorological Organization, and The Weather Channel on the "Weather Kids" campaign. The initiative used AI-generated child avatars to deliver simulated weather forecasts from the year 2050, raising awareness about the projected impacts of climate change.
Wurth Corporate Communications: German industrial supply company Wurth shifted from written to video-first internal communications using HeyGen, scaling content delivery across its global workforce.
Trivago Video Localization: Online travel company Trivago used HeyGen to halve its post-production time and localize television advertisements across 30 markets.
STUDIO 47 Newsroom: German regional news broadcaster STUDIO 47 integrated HeyGen into its newsroom operations, reporting 80% faster production times and 60% cost savings.
HeyGen's technology stack combines several areas of AI research:
Diffusion Models: Avatar IV uses a diffusion-inspired audio-to-expression engine that converts speech input into photorealistic facial animations. This approach enables the system to capture micro-expressions and emotional nuances that earlier frame-by-frame animation techniques could not reproduce.
Voice Synthesis and Cloning: The platform uses neural network-based voice synthesis to clone speaker voices from short audio samples. The cloned voices maintain the original speaker's tonal qualities across different languages.
Computer Vision: HeyGen's lip-sync technology relies on computer vision models that track facial landmarks and map synthesized speech onto video frames. The system handles occlusions, profile angles, and rapid head movements.
Natural Language Processing: The Video Translate feature uses NLP models for transcription, translation, and context-aware dubbing. The system considers visual and emotional context when selecting translation styles.
HeyGen operates in the growing AI video generation market alongside several notable competitors.
Synthesia is a London-based AI video platform founded in 2017. It is HeyGen's most direct competitor in the enterprise AI avatar video space. Synthesia has raised over $536 million in total funding across seven rounds, with its October 2025 Series E round of $200 million valuing the company at $4 billion. Synthesia crossed $100 million ARR in April 2025, roughly six months before HeyGen reached the same milestone.
| Feature | HeyGen | Synthesia |
|---|---|---|
| Founded | 2020 | 2017 |
| Headquarters | Los Angeles, USA | London, UK |
| Total Funding | ~$69 million | ~$536 million |
| Valuation | $500 million (2024) | $4 billion (2025) |
| Stock Avatars | 1,100+ | 230+ |
| Languages | 175+ | 140+ |
| ARR ($100M) | October 2025 | April 2025 |
| Fortune 100 Adoption | Not disclosed | 90%+ |
| Starting Price | $29/month (Creator) | $18/month (Starter) |
| Enterprise Focus | Strong | Very strong (SOC 2 Type II, SSO, SCIM) |
Synthesia has a larger enterprise footprint, with adoption by over 90% of Fortune 100 companies and stronger compliance certifications (SOC 2 Type II). HeyGen counters with a larger avatar library, broader language support, and the technically advanced Avatar IV engine.
D-ID is an Israeli AI video company founded in 2017 by Gil Perry, Sella Blondheim, and Eliran Kuta. Based in Tel Aviv, D-ID raised approximately $48 million in total funding before being acquired by simpleshow in September 2025 for a reported $60 million. D-ID initially focused on de-identification technology (removing identifying features from photos and video) before pivoting to AI avatar creation.
D-ID differentiated itself from HeyGen by pivoting toward conversational AI, launching its AI Agents 2.0 platform that enables real-time face-to-face conversations with digital assistants. This strategic direction earned D-ID a CES 2026 Innovation Award. However, D-ID's avatar library and language support are smaller than those offered by HeyGen or Synthesia.
| Feature | HeyGen | D-ID |
|---|---|---|
| Founded | 2020 | 2017 |
| Headquarters | Los Angeles, USA | Tel Aviv, Israel |
| Total Funding | ~$69 million | ~$48 million |
| Primary Focus | Avatar video creation + translation | Conversational AI agents |
| Avatar Library | 1,100+ | Smaller |
| Starting Price | $29/month | $6/month |
| Status (2025) | Independent, profitable | Acquired by simpleshow |
The broader AI video generation market includes several other notable players:
HeyGen maintains a dedicated Trust and Safety team led by Lavanya Poreddy, who joined the company in 2023 with experience from Meta and Match Group.
The platform employs machine learning-powered scanning tools to detect potentially harmful content during upload and generation. Flagged content undergoes manual review by trained moderators. Violations may result in content removal, generation blocking, account suspension or termination, or reporting to authorities.
HeyGen requires that AI-generated avatars be created only with the expressed consent of the depicted individual. Security safeguards include live video consent verification, dynamic verbal passcodes, and rapid human review of all avatar verification submissions.
HeyGen prohibits the creation of political or election-related content on its platform. Political actors are banned from using HeyGen to create avatars for any purpose.
HeyGen is a member of the Content Authenticity Initiative (CAI), a coalition of media companies, technology firms, NGOs, and academic institutions working to establish industry standards for content authenticity and provenance.