OpenProtein.AI Launches Accessible AI Platform for Protein Engineering

A new computational platform aims to bridge the gap between advanced artificial intelligence and life science researchers. OpenProtein.AI has introduced a computational platform designed to provide scientists with access to artificial intelligence (AI) foundation models for protein engineering. The platform is offered free of charge to academic researchers and is also available to pharmaceutical and biotechnology companies.

The company was founded by individuals with backgrounds at the Massachusetts Institute of Technology (MIT), with a goal of broadening access to cutting-edge tools for biologists.

Platform Overview: A No-Code Interface for Protein Science

The platform is centered around a web interface that allows researchers to upload data and conduct protein engineering work without requiring coding knowledge. It also provides application programming interfaces (APIs) for programmatic access.

A core component is the company's internally developed protein language model, PoET (Protein Evolutionary Transformer). According to the company, PoET was trained on groups of related proteins to generate sets of similar protein sequences and can incorporate new experimental data without requiring a full retraining of the model.

The company has since released an updated version, PoET-2. OpenProtein.AI states that PoET-2 outperforms larger AI models while using fewer computing resources and less experimental data.

Founders' Vision: Democratizing AI for Biology

OpenProtein.AI was founded by Tristan Bepler, who earned a PhD from MIT in 2020, and former MIT associate professor Tim Lu. Bepler's doctoral research involved analyzing evolutionary data to predict protein sequences, contributing to early generative AI models for protein design.

Company statements indicate the platform was created to address a perceived disconnect between advanced computational tools and biologists who may not have machine learning expertise.

• Tristan Bepler stated the AI models can make protein engineering more efficient, potentially shortening development cycles for therapeutics and industrial applications. He described a broader company goal of "creating a language for describing biological systems."
• Tim Lu emphasized the importance of creating open ecosystems around AI and biology, expressing a concern that AI resources could become so concentrated that the average researcher cannot access them.

Commercial Application: Collaboration with Boehringer Ingelheim

The platform's commercial potential is being realized through a key partnership. The pharmaceutical company Boehringer Ingelheim began using OpenProtein's platform in early 2025.

The two companies later announced an expanded collaboration to embed OpenProtein's platform and models into Boehringer Ingelheim's research.

This collaboration focuses on engineering proteins to treat conditions including cancer, autoimmune diseases, and inflammatory conditions.

Future Directions

The founders have outlined several areas for future research and development:

• Working on models that can predict and design dynamic protein functions.
• Continuing research on how to describe proteins and incorporate evolutionary constraints into its models.
• Considering applications of these AI approaches to non-protein biological modalities.