MIT Founders Launch No-Code Platform to Give Scientists Access to AI Protein Models
OpenProtein.AI, founded by MIT PhD Tristan Bepler and former MIT associate professor Tim Lu, offers researchers a web-based platform to design proteins and train machine-learning models without writing code. The company provides free access to academic scientists and sells its tools to pharmaceutical and biotech companies.
Most biologists lack machine-learning expertise. OpenProtein solves this gap by wrapping complex AI models in an intuitive interface where scientists can upload data, generate protein sequences, and validate results through predictive models - all without touching code.
How the Platform Works
The platform centers on PoET (Protein Evolutionary Transformer), OpenProtein's flagship protein language model trained on evolutionary data. PoET generates sets of related proteins and learns evolutionary constraints without requiring retraining when researchers add new experimental data.
Scientists upload proteins of interest. The models generate new variants with similar properties. Researchers can then filter candidates computationally before running expensive lab experiments, reducing both time and cost.
The platform also includes APIs for researchers who prefer to code. "It's basically a no-code front-end, but we also have APIs for people who want to access it with code," Bepler said.
Last year, OpenProtein released PoET-2, which outperforms much larger models while using a fraction of the computing resources and experimental data.
From PhD Research to Commercial Product
Bepler arrived at MIT in 2014 to study computational biology under Bonnie Berger. He became interested in predicting amino acid chains by analyzing evolutionary data - work that preceded Google's AlphaFold by years.
After earning his PhD in 2020, Bepler joined Lu's lab at MIT's Department of Biological Engineering as a postdoc. Lu saw a problem: biologists wanted access to cutting-edge AI tools but lacked the technical skills to use them. The two co-founded OpenProtein to close that gap.
"We wanted to build something that was user friendly because machine-learning ideas are kind of esoteric," Bepler said. "They require implementation, GPUs, fine-tuning, designing libraries of sequences. Especially at that time, it was a lot for biologists to learn."
Early Adoption in Pharma
Boehringer Ingelheim, a major pharmaceutical company, began using OpenProtein's platform in early 2025. The companies recently announced an expanded partnership embedding OpenProtein's models into Boehringer Ingelheim's work on cancer, autoimmune, and inflammatory disease treatments.
Broader Applications Ahead
Bepler sees protein language models as a foundation for describing biological systems more broadly. "The big picture is we're creating a language for describing biological systems," he said.
Lu is focused on designing proteins with dynamic properties - those that engage multiple biological mechanisms simultaneously or change function after binding. Current experimental tools become limiting as protein engineering grows more complex.
"There's a risk that AI resources could get so concentrated that the average researcher can't use them," Lu said. "Open access is super important for the scientific field to make progress."
Learn more about AI for Science & Research and Generative AI and LLM applications.
Your membership also unlocks:
