OKcontract Labs

We research and build infrastructure for safer onchain systems, including foundational open-source libraries such as Cells. We focus on verifiable operation intents, policy-constrained execution and formal methods to create resilient software that remains understandable when an interface, credential, dependency, or operator fails.

Discover Chainwall

How we build safer onchain systems

Security claims should be explicit, testable, and independently verifiable. We design systems that remain constrained when an interface, credential, dependency, or operator fails.

A working method

Security is not a feature added at the end. We begin by assuming that one layer will eventually be compromised, then design the surrounding system so that a local failure cannot silently become total control.

That leads us to an allowed world: make identities, targets, policies, and intended operations explicit; bind approval to exact effects; and reject what has not been described. Human-readable specifications keep intent reviewable, while deterministic execution keeps machines inside the same boundary.

Transparency is part of this method. We prefer compact primitives, reproducible behavior, and evidence that others can inspect, from ordinary regression tests to symbolic proofs for the invariants that matter.

Onchain first

Put critical policy and execution in deterministic, independently verifiable systems. Minimize the hidden services and privileged operators that users must trust.

Censorship resistance

Avoid privileged chokepoints. The systems we build should remain inspectable and usable under pressure, without one vendor or interface becoming the final authority.

Formal verification

Test examples, then challenge the invariants themselves. Symbolic execution and machine-checked properties give important security claims stronger evidence than a passing happy path.

AI-ready

Human-readable, machine-executable specifications give people and agents the same explicit boundary. Automation should become more capable without gaining ambient authority.

Technology spotlight

Cells

A reactive state graph for TypeScript applications.

Cells is a small, framework-independent reactive runtime inspired by spreadsheets. Declare how values depend on one another and let the graph keep application state coherent across asynchronous work.

Built at OKcontract Labs and released as open source, Cells models dependency relationships directly instead of synchronizing copies of derived state by hand.

Explore Cells on GitHub

graph.ts

value cell

map cell

map([a, b], (a, b) => a + b)

Async is native

Promises, errors, unavailable values, and dynamic pointers are part of the value model, not orchestration bolted on around it.

Stable by design

Ranked computations keep older async results from replacing newer state, while subscribers receive settled propagation batches.

A smaller trust surface

A focused runtime and deliberately lean toolchain keep the dependency graph and its software supply chain understandable.

Vitalik Buterin presenting at Ethereum Community Conference

A security-first approach

“A more specialized and more focused way to secure onchain interactions.”

Vitalik Buterin

Supported by