KeyAI Research · Independent research

Mathematics with
inspectable evidence.

We study elliptic-curve cryptography, formalize mathematics in Lean, and build tools that connect AI research to evidence other people can check.

307ECDLP ledger rows
~268Distinct ECDLP results
172Separate ResearchOS ledger rows

Each entry has a source and a scope.
How to read the evidence

Selected results

A few concrete places to begin.

These examples explain what a checked statement gives you, where it stops, and how to inspect it.

Browse the complete result index →

Our research

Difficult questions.
Explicit boundaries.

Our cryptographic program studies the discrete-logarithm problem on secp256k1, including algebraic structure, proposed routes, and the evidence needed to evaluate them.

We also develop reusable formal analysis and number theory in a separate ResearchOS ledger.

The repository does not solve ECDLP. No secp256k1 break or proof of the Riemann Hypothesis is claimed.

Explore the research directions →

Research OS · In development

Keep the work.
Keep the reasons.

A research project needs more than its latest successful proof. It needs the sources, failed attempts, assumptions, and decisions that explain what to do next.

Research OS connects those records in a public reference deployment. A hosted multi-project product is not yet available.

For researchers and technical collaborators

Bring a question.
Help check the work.

We welcome proof review, reproducibility work, and conversations with teams building AI for mathematics. We are also recruiting a formal-research team to test the Research OS workflow.

Find a way to collaborate