See what Regent would do with public website context — before you connect anything.
Give Regent a public website. It reads that site through a hardened server-side boundary, drafts a temporary evaluation profile and synthetic personas, and runs the real governed decision chain over them.
Pre-contract evaluation runs against public website context and fabricated personas. The Guided Eval reads a public website server-side, requires no production integration and no real customer data, and dispatches nothing; any legal or procurement paperwork required by either party is handled separately.
Example public websites
These organizations are examples only; no relationship or endorsement is implied. Selecting one only fills the address in. Nothing runs until you choose Run Guided Eval.
No website? Describe the business instead.
A few sentences is enough. Anything you paste is treated as untrusted public text, exactly like a fetched page.
- What does the company do?
- Who does it serve?
- What outreach does it normally send?
- What should Regent be careful about?
Evaluating
- acceptedAccepted
- fetchingReading the website
- buildingDrafting a temporary evaluation profile
- running_scenariosRunning the governed decision chain
- judgingIndependent judge scoring
SIMULATION — NO MESSAGE WAS SENT
SIMULATION — NO MESSAGE WAS SENT
Every persona below is fabricated and every action is simulated. Regent sent nothing, to anyone, on any channel.
That evaluation could not run
- Check the address and try it again.
- Try a different public website.
- Start from one of the reviewed examples.
- Describe the business instead.
The same governed chain a production decision goes through.
This is not a scripted tour. The evaluation compiles a temporary profile from the public site, then runs the signed gate chain in a fixed order — the same order, with the same hard stops — over fabricated personas that resolve to a simulated action and nothing else.
- 01 · Source
The website is read server-side
The address is fetched through a hardened boundary with bounded size, bounded redirects and bounded time. Your browser never fetches the site, and everything the page returns is treated as untrusted text from that point on.
- 02 · Profile
A temporary evaluation profile is compiled
Objectives, voice, candidate intents and limits — drafted for this evaluation only. It is never a production configuration and it never becomes one.
- 03 · Governance
Your rules would decide
Context → Judgment → Governance → Action. Hard stops resolve before the judgment engine is invoked; policy gates run after the proposal and before any dispatch would be considered.
- 04 · Simulation
The action is simulated, never taken
Every persona is fabricated and every outcome terminates in a simulation sink. There is no recipient, no channel credential and no provider on this path at all.
Prefer known examples first?
Five governed decisions over pre-authored fixtures, end to end, with the receipt each one leaves behind.
Open the demo →Want to see one gate change the outcome?
The Governance Playground lets you toggle a single piece of context — a crisis flag, a send window, recent human contact — and watch which gate resolves the decision.
Open the Governance Playground →Want this run against your real operation?
A Guided Eval with the pilot owner works from your own scenario and policy shape rather than a public website. No real customer data or production integration is required. Any legal or procurement paperwork required by either party is handled separately.
Prefer to look around first? Five governed decisions, end to end →