Medra AI
About the Project
Medra's Physical AI Scientist platform lets scientists hand off lab work entirely: robots run and optimize protocols end to end.
We partnered with Medra's backend engineering team to build and integrate the natural-language interface at the core of that platform: scientists type instructions in plain English and the system translates them into precise robotic actions across pipetting, heating, mixing and more.
Client
Medra AI
Industry
Scientific Automation Healthcare / Robotics
Services
AI/LLM Integration
Agentic Systems
Web development
UX / UI Design
The Challenge
Scientists lacked a way to automate lab workflows without programming knowledge: every protocol change meant waiting on an engineer. The harder problem underneath: translating natural-language instructions into safe, precise robotic actions in real time, which meant integrating deeply with Medra's backend and instrument control systems.
The Process
We worked directly alongside Medra's backend team to design, prototype, build and integrate the natural-language interface into their robotic control system: Text to Protocol. We owned both the UI and the integration layer: translating natural-language instructions into the precise robotic actions executed by the platform.
Built using Torii's Antifragile system. A custom set of AI skills, commands, and rules that let our team ship consistent, high-quality UX and code alongside Medra's engineers.
The Results
Reduced experiment setup time from days to minutes. Non-technical scientists now operate robots independently, accelerating research timelines 3x.
Medra has since raised a $52M Series A and partnered with Genentech, one of the world's leading biotechnology companies, to bring AI-driven experimentation into production-scale drug discovery. All built on the natural-language platform we helped integrate.
Other Projects
