HeyGen turns text into avatar-led video and allows existing footage to be adapted into different languages. Available online at heygen.com, its platform brings together several generative AI components in an interface designed for professionals, including those with no video editing skills. Its aim is to shorten the audiovisual production process, from script to distribution, for content in which a person speaks directly to the viewer.
From early avatars to a platform for businesses
Founded in 2020 by Joshua Xu and Wayne Liang, the company initially operated under the name Movio before adopting HeyGen in 2023. It belongs to a generation of companies applying advances in machine learning to facial representation, voice and lip synchronization. Rather than focusing on generating cinematic scenes, it prioritizes presenter-led video.
In 2024, HeyGen announced a funding round led by Benchmark to support its growth. Its business model relies on subscriptions, with features and production capabilities varying by plan, as well as packages designed for organizations. It operates in a competitive market that includes Synthesia and several specialists in video or synthetic voice.
Creating, translating and personalizing videos
The service lets users choose an avatar from a catalog or create a personalized digital representation from recordings. Users provide a script, select a voice and compose their video using visual elements. Intended uses include training modules, sales presentations, product explanations and internal communications. The benefits are particularly tangible when this content needs to be updated regularly.
Video translation is another major focus. HeyGen combines translation, voice generation and lip movement adaptation to render a spoken presentation in multiple languages. This automation does not eliminate the need for review: industry terminology, pronunciation, cultural nuances and fidelity to the message still need to be checked. Application programming interfaces also allow certain functions to be integrated into clients’ tools and production processes.
What next?
HeyGen’s trajectory will depend as much on the quality of its output as on its integration into business practices. Interactive avatars, in particular, open up possibilities for assistance, learning and personalized exchanges. Their usefulness, however, will need to be assessed in terms of response relevance, latency and the acceptability of these synthetic conversational partners.
Managing risks is the other central challenge. Reproducing a face or voice raises questions about consent, data protection and identity theft. For its European clients, transparency about artificial content and regulatory requirements will influence technology purchasing decisions. HeyGen will therefore need to combine ease of creation, control over usage and safeguards suited to sensitive content.