Lemon Slice Nabs $10.5M from YC and Matrix to Build Out Its Digital Avatar Tech
Lemon Slice has raised $10.5M in seed funding to develop its digital avatar technology, allowing businesses to create interactive avatars from a single image. The new Lemon Slice-2 model aims to enhance AI agents by adding a video layer, with applications across education, e-commerce, and more.
Developers and companies are increasingly integrating AI agents and chatbots into their apps, but so far, most of these have been limited to text-based interactions. Digital avatar generation company Lemon Slice is developing a video layer for these chats using a new diffusion model that can generate digital avatars from a single image.
The model, called Lemon Slice-2, can create a digital avatar that operates on top of a knowledge base to fulfil any role required of an AI agent, including answering customer queries, assisting with homework questions, or providing mental health support.
"In the early days of GenAI, my co-founders started to play around with different video models, and it became obvious to us that video was going to be interactive. The compelling part about tools like ChatGPT was that they were interactive, and we want video to have that layer," said co-founder Lina Colucci.
Lemon Slice claims the model is a 20-billion-parameter system that can run on a single GPU to livestream videos at 20 frames per second. The company is making the model available through an API and an embeddable widget that companies can integrate into their sites with a single line of code. After creating an avatar, users can change the character's background, styling, and appearance at any time.
In addition to human-like avatars, the company focuses on creating non-human characters to meet diverse needs. Lemon Slice is using ElevenLabs' technology to develop the avatars' voices.
Founded in 2024 by Lina Colucci, Sidney Primas, and Andrew Weitz, Lemon Slice is betting that its general-purpose diffusion model for avatar creation will differentiate it from competitors.
"The existing avatar solutions I've seen to date add negative value to the product," Colucci stated. "They are creepy, and they are stiff. They look good for a few seconds, and as soon as you start interacting with them, it feels very uncanny, and it doesn't put you at ease. The thing that has prevented avatars from really taking off is that they haven't been good enough."
To fund the expansion of its efforts, the company announced it has raised $10.5 million in seed funding from Matrix Partners, Y Combinator, Dropbox CTO Arash Ferdowsi, Twitch CEO Emmett Shear, and The Chainsmokers.
Lemon Slice has implemented guardrails to prevent unauthorized face or voice cloning and uses large language models for content moderation.
While the company did not disclose the names of organizations using its technology, it noted that the model is being used in areas such as education, language learning, e-commerce, and corporate training.
The startup faces intense competition from video generation companies such as D-ID, HeyGen, and Synthesia, as well as other digital avatar creators like Genies, Soul Machines, Praktika, and AvatarOS.
Ilya Sukhar, a partner at Matrix, believes avatars will have significant value in areas where video is prominent, such as learning from YouTube rather than reading long-form text. He emphasized that Lemon Slice's technical expertise would give it an edge over competitors.
"It's a deeply technical team with a track record of shipping ML products, not just demos and research. Many of the other players are bespoke to particular scenarios or verticals, and Lemon Slice is taking the generalized 'bitter lesson' scaling approach (of data and compute) that has worked in other AI modalities," Sukhar said.
Y Combinator's Jared Friedman also believes that Lemon Slice's diffusion-style model enables the creation of any avatar, unlike some other startups that focus either on human-like avatars or game-like characters.
"Lemon Slice is, I believe, the only company taking the fundamental ML approach that can eventually overcome the uncanny valley and break the avatar Turing test. They train the same type of model as Veo3 or Sora: a video diffusion transformer. Because it is a general-purpose model that does the whole thing end-to-end, it has no ceiling on how good it can get; the others top out below photorealistic. It also works for both human and non-human faces and requires only an image to add a new face," Friedman said.
Currently, Lemon Slice has eight employees, and the company plans to use the raised funds to hire engineering and go-to-market staff, as well as to cover the computing costs for training its models.
What's Your Reaction?
Like
0
Dislike
0
Love
0
Funny
0
Angry
0
Sad
0
Wow
0