Skip to main content

Contract IDE — the state-contract studio

The Contract IDE is where you build state contracts. A state contract is a small physics state machine that runs on your machine: you declare the states it can be in — say idle → watching → alert — and the transitions between them. Each transition carries a real operator — the verbatim equation of a physical law, NIST CODATA-bound, to ≤0.1% — and fires when a value crosses your threshold or on a time trigger. The kernel evaluates it on the 1.287 Hz heartbeat and writes every fire to your tamper-evident entangled state. It's state-first, pulse-synchronised, physics-grade, and chain-rooted.

Three ways to author

Template

Pick from a library of industry examples by name, preview it, and deploy — the fastest start.

Expert (JSON)

Write the contract directly as JSON when you want full control over states, transitions, and triggers.

Generate

Describe what you want and have the kernel architect a first draft — an optional convenience, then you edit it like any other.

Whichever path you take, the flow is the same: preview → one-click deploy. A deployed contract is saved to your machine and starts safe (paused-ready), so nothing fires until you say so.

Manage, version, roll back

Every contract you deploy lives under Saved — open, edit, pause and resume, fire manually, or change where it runs. Each deploy snapshots a version, so you can roll back at any time, and every fire is a row on your entangled-state chain, viewable in the observer. The studio is the contract's home; the machine is its record.

Run the compute on your own server

By default a contract's compute runs on your machine. But you can point a contract's execution at your own server with a signed webhook URL: the kernel sends each transition's payload to your endpoint, your code computes, and the signed result is written back to your chain. The contract, its state, and its proof stay anchored to your machine — you just host the work wherever you like. Set it per-contract under Saved → Hosting. It's the cleanest expression of the framework's split: you can own the computation, and the machine still owns the proof.

Open it