A video of a baby interviewing a dog on a podcast became viral last month.
No, it was not real. It was a video generated by the AI-Créée by the actor Jon Lajoie, who used Hedra, a platform for generation of AI videos, to make the animation.
The Hedra platform allows users to generate images, videos and audio with its content creation studio based on the web.
“Our model and technology focus on the most controllable and convincing characters, whether it is a hyperrealist human or an animal character or even an animal,” said Business Insider, CEO of Hedra, CEO, Michael Lingelbach.
HEDRA announced on Thursday that it has raised a series of fundraising of $ 32 million carried out by the Andreessen Horowitz infrastructure fund. The tour included return investors such as A16Z Speedrun, Abstract and Index Ventures.
Since its launch in 2024, the AI video startup has quickly increased capital. In August, he announced an investment cycle in seeds of $ 10 million. In March, Amazon’s Alexa Fund announced that it had invested in the startup and several other AI companies. Hera said that he had raised a total of $ 44 million but had not revealed an assessment.
Competition in the generating AI is hot, with Buzzy companies such as legends, Heygen, Synthesia and track construction technology around video and avatars (Hedra has specified that it is not an avatar company).
“We are not trying to compete with Google Veo, we are not trying to compete with Sora,” said Lingelbach. “We really focus firmly on the construction of the best character models, and it is something that, with this additional capital, we can do another function in the field.”
The characteristic “Omnimodal” model of Hedra combines images, text and audio to generate a video. The creation of a character with Hedra begins by downloading an image, then downloading audio that they have already saved (like a podcast) or generated using models of vocal text like Elevenlabs.
“The voice and the video are seeing a rapid development right now,” said Langelbach. “We made a big leap forward on the naturalness of the expression with our current model.”
The Hedra platform is also users to integrate outdoor models like Elevenlabs, Google Veo and “All in One Workflow”, said Lingelbach.
Extend beyond the creator’s economy
Hedra’s main user basis has been professional marketing creators and specialists, said Lingelbach.
“We are already seeing a massive influx of content generated by AI,” said Lingelbach. “My Instagram and Tiktok flow is filled with various memes and also more serious content now that is generated by AI.”
From comedy sketches to the contents of the faceless creator via the speaking babies, Hedra has already seen a wide range of use cases. The content of the podcast, in particular, has been a popular application of Hedra technology.
“This is not really something that we planned at the start, but it certainly led a lot of our use,” he said.
In addition to the viral trend of podcasts hosted by Bébé ai that people have created using Hedra, others used Hedra to create Ghibli-Studio style videos of the classic podcast interview clip.
With its recent increase, Hedra plans to develop in more business marketing requests, extend its team and open an office in New York.
Read the Pitch Deck Hedra used to raise its series A:
Note: Some slides have been raised in order to publicly share the game.
Hède
Hedra focuses on narration and characters.
Hède
Here is what the slide says:
Each good story is made around the characters. But so far, creating convincing characters has been the most difficult part of content creation.
The game explains the “Omnimodal Foundation model” of Hedra which allows people to quickly generate digital characters.
Hède
Here is what the slide says:
At Hedra, we have built the best performance model in the world in the world which uniquely combines video, voice, movement and emotion in an ever possible way.
The Characteristic-3 model of Hedra is the first Omnimodal Foundation model in the world in production.
The only model that supports human, animated and animal characters. And it works with any angle or framing.
Built to effectively prioritize unified models
The entire model was developed with a budget of less than $ 2 million
HEDRA customers range from everyday consumers to creators and marketing specialists.
Hède
Here is what the slide says:
We have seen three key customers segments rely on Hedra for their different characteristics of character performance.
Consumers
Consumers love the fun content they can create on Hedra – from memes to clips to short films.
Prosimation
Prosumers use HEDRA for use cases such as video podcasts, social media and IP development.
Marketing teams
Marketing teams use Hedra to create UGC, product tutorials and other content to stimulate sales conversions.
Hedra plans to develop in more business offers.
Hède
Here is what the slide says:
And we have already aroused heavy interest in companies for professional use and have signed notable customers in the first company.
Upcoming corporate features:
Team management
Guardians IP
Corporate level
The game highlights the Hedra research team and its owner technology.
Hède
Here is what the slide says:
Hedra is about to be the first company to break the “Valley Uncanny”.
We have a best class research team that allows us to train owner omnimodal models that no competitor can develop
We are a company focused on products that builds foundation models for real world applications, listen to our users and ships frequently.
We are an effective capital team that built the character 3 out of less than $ 2 million in capital and a small research team.
Then the game presents the team.
Hède
Here is what the slide says:
We have brought together the best team to own this category – marry in -depth research with the design of AI -Native products.
Key leadership team:
Michael Lingelbach: Founder / CEO
Stanford Phd Student of Fei-Fei Li and Jiajun Wu. Principal author of 3 Real-time dissemination documents. Recipient of the prestigious Stanford higher education scholarship.
Hongwei Yi: Chief of research
The former doctoral student of Michael Black, principal researcher behind the first audio broadcasting model of the video to arrive on the market in the United States.
Wei Li: Research research
CONTIBLE CONTRACTER BASIC TO Google BARD / GEMINI, PALM-2 and T5, with more than 8 years of experience at Google Brain / Deepmind.
Jason Wilson: Engineering manager
Predicted in engineering in Nava Benefits (startup of serial B of Thrive) and director of engineering at Descartes Labs.
Alan Guo: Chief of staff
MBA of the Harvard Business School. Previously, worked in growth and strategy at Disney, Jubilee Media and Firework.
Ramin Keene: Main engineer
Old CTO at Stockx and twice founder.
Hedra concludes his deck by saying: “We are just beginning”.
Hède
Here is what the slide says:
We are on a mission to build the best narrative platform from start to finish of the world. And we just start.
We have already built the world’s first performance model in the world on a budget of $ 2 million.
Prosumers, consumers and business marketing specialists love and depend on us for their content.
We are now working on even larger versions of features that will reinvent the creation workflows.