Keyframe Labs
Spring 2026 NewTurn agents into lifelike video calls. $0.06 a minute.
Our interactive video models turn agents into lifelike video calls. Deploy in minutes with a few lines of embed code. Photorealism, emotion, and scale for just $0.06 per minute.
AI Investor Summary
Keyframe Labs is building lifelike, emotionally expressive AI humans for real-time video calls, aiming to revolutionize customer service and engagement. With a strong technical founding team from Zipline, Ericsson, Google, and Meta, they are poised to capture a large market opportunity. Their compelling $0.06/minute pricing offers significant scalability, though early traction and technical defensibility require further validation.
Key Highlights
- ● Founders possess strong, complementary technical backgrounds in AI, computer vision, and real-time communication from top universities and reputable companies.
- ● Addresses a significant market opportunity with the growing demand for AI-driven customer engagement and virtual agents.
- ● Low per-minute cost ($0.06) is a compelling value proposition for scalability.
Risk Factors
- ● Lack of demonstrated revenue and user traction makes it difficult to validate product-market fit and business model viability.
- ● Technical defensibility and moat need to be clearly articulated and proven to withstand competition in the rapidly evolving AI space.
- ● Achieving consistent photorealism and genuine emotional expression in real-time video calls at scale is a significant technical challenge.
Founders
Kaahan is co-founder and CTO of Keyframe Labs, building the world's first photoreal, emotionally expressive AI humans for real-time video agents. Previously, he spent five years at Zipline as a Perception Engineer, building their camera-based detect-and-avoid system for autonomous drone operations. His work spans computer vision and applied ML, with research at CVPR and JAMA. B.A. CS, UC Berkeley.
Parth is co-founder & CEO of Keyframe Labs. Before Keyframe, he spent 5 years at Ericsson on frontier AI, founding and leading a team of 30 building foundation models for voice & video synthesis on one of the world's first petascale human datasets, shipping Ericsson's first generative AI products. Before that, as a research scientist at AT&T Labs he developed novel neural architectures for photorealistic 3D telepresence, enabling the world's first 3D call over production networks. B.S. CS UCLA.
Score Breakdown
Strong technical team with deep AI/ML expertise and relevant industry experience. Kaahan's background in perception engineering at Zipline and research at CVPR/JAMA is highly relevant. Parth's experience at Ericsson leading generative AI initiatives and AT&T Labs on 3D telepresence, coupled with his Google/Meta background, demonstrates significant product development and leadership capabilities. Both founders have strong CS education from top-tier universities. Founder-market fit seems good given their combined expertise in AI, computer vision, and real-time communication. [Boost +1: Founder from Google]
Large addressable market in customer service, sales, education, and entertainment where lifelike virtual agents can enhance user experience and engagement. The trend towards AI-powered customer interactions and the increasing sophistication of generative AI models suggest strong timing. The competitive landscape is emerging but not yet saturated, with significant room for differentiation. Regulatory tailwinds around AI adoption are generally positive, though ethical considerations around AI representation will be important.
Product shows promise with its claim of photorealism, emotion, and scale at a low cost per minute. The technical differentiation lies in turning 'agents' into lifelike video calls, implying a focus on real-time, interactive, and emotionally expressive AI humans. The defensibility/moat is not yet clearly established and will likely depend on the proprietary nature of their models and the quality of their output. UX quality is implied to be high due to 'deploy in minutes with a few lines of embed code,' suggesting ease of integration. Platform potential exists if they can build a robust API and ecosystem around their AI humans.
Early stage with $0 revenue reported. Traction is primarily based on early press coverage and inclusion in Y Combinator, indicating investor interest. Partnerships and user growth are not yet detailed, making it difficult to assess market validation. The presence on HyperStore suggests some initial product exposure, but it's not a direct revenue indicator. [Boost +2: Tier-1 VC: accel]
News
Keyframe Labs has launched real-time, generative models that turn AI into lifelike video calls, emphasizing photorealism, emotion, and scalability at a cost of $0.06 per minute.
Keyframe Labs has released Persona-1-Live, a real-time streaming variant of their Persona-1 model, enabling dynamic, low-latency conversations with human-like interactions.
Keyframe Labs was among the Y Combinator launches, noted for its real-time, generative models that turn AI into lifelike video calls.
The Keyframe Labs blog features research, ideas, and product updates on their AI-powered human computing frontier, including updates on Persona-1.
Keyframe Labs uses generative models to create photorealistic AI personas for natural video conversations, offering no-code/low-code integration and scalability for businesses.
Keyframe Labs enables the creation of photorealistic, conversational AI humans for real-time video calls, offering no-code and low-code integration with pricing as low as $0.06 per minute.
Keymakr has launched a new suite of tools and services focused on data for Large Language Model (LLM) agents and agentic AI systems, addressing the need for high-quality domain-specific datasets and human feedback pipelines.
Keyframe Labs, a Y Combinator Spring 2026 startup, provides interactive video models to transform AI agents into lifelike video calls with photorealism and emotion at a low cost.
Keyframe Labs is an organization focused on pioneering human computing, with repositories related to documentation and a framework for building real-time voice AI agents.
Keyframe Labs introduces Persona-1, an expressive, faster-than-realtime talking head model designed for real-time interactions with personas at scale.
Keyframe Labs offers interactive video models that turn AI agents into lifelike video calls with photorealism, emotion, and real-time latency, starting at $0.06 per minute.
Quick Info
- Batch
- Spring 2026
- Team Size
- 2
- Location
- Unspecified
- Founders
- 2
- Scraped
- 4/10/2026