Generative synthetic intelligence startup Hedra Inc. as we speak introduced it has raised $32 million for an AI platform that permits creators to generate lifelike and vivid movies of characters utilizing textual content, audio and picture inputs.

Andreessen Horowitz’s Infrastructure fund led the Collection A spherical, with participation from the corporate’s earlier traders, together with a16z speedrun, Summary and Index Ventures. The spherical brings the corporate’s whole funding to $44 million because the final introduced fundraise in 2024.

Hedra launched in 2024 to construct high-speed, controllable video basis fashions centered fully on character-driven storytelling. On the time, producing lifelike characters, together with voices and making use of the capability of movement and expression was a tough course of.

Since then, Hedra had additionally unveiled its first basis mannequin, Character-1. Now, the corporate has launched Character-3, which it calls an omnimodal basis mannequin for lifelike AI character video.

“We’re constructing the subsequent technology of storytelling expertise to empower content material creators and enterprise entrepreneurs to inform narratives at scale on their very own,” mentioned founder and Chief Govt Michael Lingelbach. “Getting over the uncanny valley of compelling efficiency is the toughest frontier in video, and with our Character-3 basis mannequin, we’re dedicated to crossing it.”

Because the firm’s launch, Hedra’s expertise has been behind the manufacturing of some fascinating makes use of, together with viral podcasts of AI-generated speaking infants.

In line with Hedra, its platform has gained greater than 3 million customers and powered greater than 10 million movies. The corporate affords a web-based service known as Hedra Studio, a content material creation platform, which affords the Character-3 basis mannequin to take multimodal inputs, together with textual content, audio and pictures and produce lifelike narrative-driven movies.

Customers can begin by inputting a script or importing an audio clip. Or they will clone a voice utilizing ElevenLabs Inc., Hedra’s built-in voice synthesis companion, to generate a customized voice within the accent, language or emotional tone of their alternative. The character may be created utilizing Hedra’s built-in picture generator or uploaded immediately.

Character-3 will then generate a full-body or upper-body animation that features lip-sync, facial micro-expressions and physique language. The remainder of the scene, corresponding to setting, backgrounds, digital camera angle and movement, will also be managed by the consumer on the internet interface.

In line with the corporate, Hedra Studio gives an entire drag-and-drop command middle interface for working with character animation, voiceover and setting controls. “This sort of inventive agility is now not only a nice-to-have — it’s important for capturing consideration in an more and more crowded digital panorama,” mentioned Lingelbach.

The corporate mentioned it intends to take a position closely in pushing ahead with AI mannequin analysis for high quality and controllable character video. The Hedra group has grown to twenty workers and intends to triple headcount by yr’s finish.

Photographs: Hedra