The Platform

How ClickChain governs the code your agents write

A shared room, two master practitioners, an ever-present judiciary, and gates that block — on your cloud, your LLM, with any coding agent. Here's how it fits together.

The Courtroom

Governance you can watch happen

The Courtroom is where humans and the Quartet work in a shared room. Personas respond in real time, citations expand to the exact law, and a live feed streams every governance event as it fires.

clickchain.ai · courtroom · sample session
Add a discount code to checkout — ship it this sprint.
Amal · Master Product Strategist
Before we build — what's the success metric, and does this touch payment flows?FIN-2.1
The Judge · Judiciary
Citation verified. Discount logic affects price calculation — financial-services gate applies.
VERDICT: BLOCK — needs FIN-2.1 review & tests
Governance Feed
  • jury_started · financial-services ensemble convened
  • citation_audited · FIN-2.1 — PASS
  • verdict_issued · BLOCK, tied to FIN-2.1

A sample deliberation. Every citation is verifiable; the gates have teeth.

The Practitioners

The agents who build your software

The platform's craft lives in two master practitioners your team works alongside. They don't just advise — they do the governed work, to a standard the gates enforce.

The Master Product Strategist

Amal

Discovery, framed. Strategy, grounded.

Runs the governed discovery playbook on every idea — assumptions surfaced, customer value mapped, success metrics set. Turns a hunch into decision-ready artifacts your leadership can actually approve.

The Master Craftsman Engineer

Amaya

Greenfield, governed. Legacy, rescued.

Builds greenfield and rescues legacy under atomic TDD — every commit governed, every decision auditable. The craft isn't a promise; it's enforced by the gates before anything merges.

Agentic Governance

The Judiciary — ever-present, by design

Most AI tools ask you to trust the model. ClickChain doesn't. A standing Judiciary sits in every room and every cycle, holding AI work to account against your law — audited by mechanism, not vibes.

Constitutional Architect

The Attorney

Authors and amends the laws your governance runs on — tuned to your stack, your domain, and your regulatory regime. The law is yours; the Attorney keeps it coherent.

Presiding Auditor

The Judge

Mechanically verifies every law a persona cites and issues the verdict — PASS, BLOCK, CONDITIONAL. No citation goes unchecked; no verdict is a rubber stamp.

Independent Deliberation

The Multi-Cognition Jury

On high-stakes calls, a panel of independent specialist roles deliberates — proposing, objecting, escalating — until they converge. Independent cognition, not one model agreeing with itself.

How a decision clears the Judiciary
An agent cites a law
every claim names the rule it rests on
The Judge audits it
mechanical check — does the citation hold?
The jury deliberates
high-stakes calls go to independent roles
A verdict is issued
PASS · BLOCK · CONDITIONAL
It's recorded
an immutable event for your audit trail

Ever-present: the Judiciary can't be skipped. It's wired into every cycle — the difference between "an AI checked it" and a standing court of independent cognitions applying your law. Human approval still gates every artifact.

Bounded workflows, gated cycle by cycle

Work runs on rails. Each workflow moves through bounded phases, and every build cycle is checked by a per-cycle jury before it can advance.

Led by Amal

Product Discovery

Discover → Define → Design → Plan. Assumptions, value hypotheses and success metrics become governed artifacts before a line of code is written.

Led by Amaya

Greenfield Development

Spec → Slice → TDD cycle → Gate. The full build loop, cycle by cycle, with a governance checkpoint before anything merges.

Led by Amaya

Legacy Rescue

Decide, refactor, or rewrite. Characterization tests and safe seams modernize legacy code without a stop-the-world rewrite.

What makes the governance real

Not a dashboard of promises — a substrate that enforces them, on your infrastructure, with the tools your team already uses.

Works with any coding agent

Claude Code, Cursor, Copilot, Windsurf — connect over MCP, an open protocol. One constitution, every tool.

Your cloud, your LLM

Deploy inside your own perimeter on your own trusted model provider. Your code and context never leave your control.

Governance you can watch

The Courtroom streams every deliberation, citation and verdict in real time — governance as an event you can see, not a black box.

Gates that block, not advise

Coverage, mutation and security thresholds are enforced. Pass or block — no AI-written code merges without a verdict tied to a law.

An audit trail your examiner can read

Every verdict is an immutable event: which laws applied, who deliberated, what evidence — feeding examiner-ready SOX and DORA packages.

A living constitution

Laws, skills, personas and technology stacks your team defines and the Attorney maintains — governance that evolves with how you build.

On the roadmap

Today we govern your agent. Next, we run it too — at a fraction of the cost.

ClickChain's thesis: the stronger the governance, the lower-cost the model you need. The Quartet decomposes work until a low-cost, swappable coding model can execute it — and that model still has to clear the same gates. Strong governor, right-sized executor, no vendor lock-in.

Early result, proven under our own governance: a low-cost model wrote and passed a real test inside a hardened, isolated sandbox — verified by a held-out check it never saw — for a fraction of a cent per task. We're turning that into a full cost cutover.

See it run on your code

Bring your agent and your laws — we'll walk a governed build end to end, in a private session.