A presenter generated from a script
Synthesia builds video software in which a script drives an avatar's spoken delivery and visible performance. Its EXPRESS-1 work connected the content of the script with tone, facial movement and lip synchronization. This places the company between generative-model research and the practical task of producing narrated business video. [1]
The company describes these changes as generated performances. They do not mean that a synthetic presenter understands or endorses the words it appears to say. The current stock-avatar documentation also distinguishes avatar generations and identifies Express-1 avatars as built from professionally filmed actor footage. [1][3]
Expressive Avatars reach a launch event
Synthesia's official product notice scheduled the Expressive Avatars launch for April 25, 2024, with a hands-on test for attendees. The feature generated expressions and head movements from the script instead of asking an editor to animate every response manually. [2]
This is a different control problem from general cinematic video generation: the output centers on a continuing presenter delivering specific material. Current documentation makes Express-1 stock avatars available through the video API, which lets software select an avatar by identifier and produce video programmatically. [3]
Sources
- Expressive Avatars powered by Synthesia’s new EXPRESS-1 model are here
Synthesia · retrieved Oct 23, 2026
- 😍 Event: Introducing Expressive AI Avatars
Synthesia · Apr 24, 2024
- What stock avatars are available in Synthesia?
Synthesia · retrieved Oct 23, 2026