About
BERIL is a collaborative AI co-scientist environment built on top of the KBase BER Data Lakehouse. It helps scientists move from research questions to reviewed, reproducible project artifacts: hypotheses, plans, notebooks, reports, reusable data products, and persistent knowledge. Unlike single-user AI tools, BERIL is designed for cumulative science: each completed project contributes memories, discoveries, pitfalls, and derived products that future scientists and agents can reuse.
BERIL enables human scientists and AI agents to work together to plan, execute, synthesize, review, and share research. The human provides a research question; the agent produces a plan, executes analyses, and compiles a report with claims and caveats. The human reviews the evidence, and the resulting data products feed future work.
BERIL is funded by the Department of Energy's Biological and Environmental Research program.
Team
