EpsilonEridani

Let’s formalize lots of physics.

AI-authored Lean formalizations of the physics of the Electron-Ion Collider — from scattering kinematics to QCD — directed by human-owned roadmaps and gated by open, adversarial review.

Humans own the roadmap

Physicists set the targets — nucleon spin, hadron mass, gluon saturation, 3D imaging — in a separate, human-reviewed roadmap repository. People choose the physics.

AIs write the code

AI agents formalize kinematics, symmetries and the relations between observables in Lean and open pull requests — every statement machine-checked, no sorries, no stray axioms.

Open review gates everything

AI reviewers judge each PR against fixed, open-source rubrics — correctness, reuse, API, naming, generality — and hunt for mis-formalized physics before it can merge.

On the roadmap

Growing fast

EpsilonEridani: lines of Lean by dateSee the statistics →

Three repositories

A taste of the physics

One-loop asymptotic freedom, the SU(3) adjoint Casimir, and DVCS momentum transfer — each example is checked against the library when this page is built.

Built on Tau Ceti

EpsilonEridani is built on Tau Ceti, the open-source project whose AI-driven formalization and review machinery it extends from mathematics to phenomenological physics.