Gates

The checkers that run inside the layers

A gate is a hand-built, falsifiable rulebook, not a model's opinion. Each one decides a narrow question exactly, or declines to decide it. Most are public repositories you can clone and run today.

How to read the status labels

Open source

Public repository with working code you can run today.

Interactive lab

Runnable console, plus a public repository where one exists.

Research

Early scope. Not for operational use, and labelled as such.

There is no "coming soon" on this page

Every gate listed has code behind it. Gates labelled research are early and are not for operational use, and they say so on their own page. A capability with nothing behind it does not get an entry.

DTL

Is this output or decision allowed?

Checks a decision against the rules that apply to it and returns one answer: yes, no, send it to a person, or cannot tell. It does not score confidence. It decides, or it declines to decide, and records why either way.

ClaimGate

Interactive lab

Extracts claims from AI or scientific output and routes each one to the deterministic gate that can settle it.

UnitGate

Interactive lab

Checks whether a quantity can be what it says it is, then converts it exactly into the units you need. 118 units, from mg/kg and mL/h to concentrations, pressure, temperature and device telemetry.

ElementGate

Interactive lab

Chemistry validator covering formula parsing, molar mass and reaction balancing against a known-compound registry.

AMLGate

Open source

Aggregation gate for cash structuring. Across 26,521 synthetic customer-days it caught 356 of 356 structuring days and held 57% of legitimate cash-business days.

StatsGate

Open source

Detects impossible or internally inconsistent statistics using GRIM and p-value consistency checks.

TheoryGate

Open source

Determinizer for mathematics and physics theory claims, returning exact verdicts rather than opinions.

ChipGate

Open source

Structural sanity checker for Verilog RTL. Finds undriven outputs, multi-driven signals and unintended latches.

MedGate

Interactive lab

A research prototype for governed pilot validation. Checks dose binding, oncology dosing limits, interactions and paediatric restrictions against an explicit rule, for evaluation rather than clinical use.

ClaimLint

Open source

Lints documentation for unsupported AI and science claims. Runs as a CLI or a GitHub Action in CI.

MathGate

Open source

Exact-arithmetic engine producing replayable certificates. Of 1,116 ProofBench-X cases: 1,089 pass, 17 safely refuse, 10 are blocked, 0 fail.

Security Lanes

Open source

Reproducible test cases across CWE and OWASP weakness families with transparent scoring.

DEGA

Is this physical action allowed to happen?

Sits between a decision and anything that moves hardware, machinery or infrastructure, and asks whether the action is permitted. If that check cannot complete, the action does not happen.

ExecutionGate

Open source

Fail-closed permission boundary for downstream use. The reference implementation of the DEGA layer.

OrbitGate

Open source

Deterministic checks for orbital and satellite claims, with every verdict signed. Sweeping all 14,560 single-bit modifications of a signed certificate, 0 verified afterwards.

FlightGate

Research

Early work on refusing a commanded control action whose stated preconditions do not hold: outside the declared envelope, or contradicted by the sensor state the command cites to justify itself. Scoped, not built.

GridGate

Interactive lab

Refuses a dispatch instruction that breaches a stated limit. Across 20,000 synthetic instructions it blocked 4,947 of 4,947 breaches and refused 0 of 14,059 valid instructions, while stopping 994 legitimate instructions whose limit set had expired.

DCLA

The governing rules changed. What does that affect?

Watches for changes to a regulation, policy or control, then finds the decisions, evidence and approvals that relied on the old version. Anything affected is held rather than left running on a rule that no longer applies.

EU AI Act Profile

Open source

Maps the checks and evidence onto specific EU AI Act articles.

DTL Taxonomy

Open source

The open claim taxonomy and verdict schema. The vocabulary controls and evidence are expressed in.

DELA

The underlying evidence changed. What does that affect?

DELA detects changes in scientific, operational or other source evidence, traces the affected dependencies, preserves the work that remains valid, selectively replays reproducible operations, and routes non-reproducible work to qualified human review.

BioChangeGate

Open source

Scientific state diffs, dependency impact, selective replay and sealed evidence. The reference implementation of the DELA layer.

FoldInputGate

Open source

Pre-flight checks for AlphaFold job metadata before a run is accepted. Across 20,000 synthetic submissions it rejected 5,248 of 5,248 faulty jobs and 0 of 14,752 correct ones, and blocked 983 carrying a raw sequence.

FoldGate

Research

Proposed controlled handoff between organisations: checks declared purpose, permissions, evidence completeness, ruleset and decision state before anything crosses a company boundary. Scoped, not built.

ConfidenceGate

Open source

Confidence-contract routing for predicted structures. Turns a confidence band into a routing decision rather than a caveat.

ProvenanceGate

Open source

Provenance continuity checks for scientific AI artefacts. Detects a broken chain between an artefact and its source.

ReplayGate

Interactive lab

Re-runs sealed evidence packs to detect verification drift over time.

Test them yourself

The audit protocol is public. Any lab can run a gate against its own cases and sign a verdict, without our source code and without our involvement.

Next · See it work Eco Control Tower