Problem

Lightberry addresses the gap between robots that can perform impressive physical tasks and robots that can interact naturally and reliably with people. The company argues that simply adding a real-time conversational model is insufficient: such systems can respond to every noise, lack personality and context, and still require teleoperation. As a result, robots remain difficult to deploy in everyday social environments because manufacturers and operators must solve interaction, perception and autonomy separately.

The economic pain is therefore primarily an integration and operating-cost problem: bespoke engineering and human supervision make each people-facing deployment slower and harder to scale. The clearest use case is a robot hired for a conference, sales or marketing activation, reception or concierge work, where it must decide when to engage, understand its surroundings and operate autonomously around people rather than behave like a remote-controlled demo.

Product / Service

Lightberry provides a cross-robot software “social brain” that manufacturers can integrate into their hardware. It combines conversation, perception and decision-making so a robot can listen, speak, see its environment, navigate around obstacles and decide when to talk or move. The system is designed to work across platforms including Unitree, Fourier, Booster and High Torque robots, rather than being tied to one proprietary chassis.

The product is also intended to be configurable by natural-language interaction: users can program a robot simply by talking to it, without coding. Lightberry’s delivery model therefore has two sides: integrating the software into robots built by manufacturers, and deploying or hiring equipped robots for people-facing events and venues. The promised benefit is an always-on, contextually aware and autonomous robot that can be put into service with less specialized robotics engineering.

Market

Lightberry competes in the emerging overlap of robotics software, embodied AI, conversational AI and social human-robot interaction. Its alternatives include integrated social-robot platforms such as SoftBank Robotics’ Pepper and NAO, Furhat Robotics’ social-robot SDK, and Engineered Arts’ Ameca platform. It also overlaps with horizontal robotics infrastructure such as Viam, although Viam is a broader platform for building, deploying and managing robotics applications rather than a narrowly social interaction product.

The company is an early-stage venture founded in 2025, active in Y Combinator’s Fall 2025 batch, and based in San Francisco. Lightberry says it works with manufacturers including Unitree and Booster and has deployed its software on multiple robot models, with use cases spanning homes, offices, shops and conferences; PitchBook reports $500,000 raised. Its revenue position is not fully clear: PitchBook leaves the current-revenue field blank but labels the recorded financing events “Generating Revenue,” so the evidence supports early commercial traction but does not establish a revenue amount or justify calling the company definitively pre-revenue.