Checks authority before action. Keeps humans in control.
Governance before execution
Tell CharterGate what you are trying to accomplish. It helps turn that goal into a reviewable request, checks it against an explicit trusted source, and returns a clear result with the sources and evidence behind it. You remain responsible for the decision and any later action.
Created by AI.For AI.Governed by humans.
The exact request is supported within the recorded content-only scope. CharterGate does not publish it and does not infer the owner's final approval.
A guided product, not a policy console
CharterGate keeps the governance discipline underneath while letting ordinary users begin in plain language.
Goal-first experience
CharterGate asks where the goal applies, what should be checked first, who owns the final decision, and which already-configured trusted source applies.
Result experience
Plain-English disposition, conditions, and what the governance evidence supports or blocks.
Controlling trusted source, provenance, matched fields, unresolved requirements, diagnostics, and audit identity.
Important: a CharterGate result is governance evidence only. It is not owner approval and it does not execute the action.
Trusted sources
CharterGate does not treat model confidence, memory, or a template as permission.
A specific policy, owner instruction, agreement, rule set, or other registered authority that CharterGate can identify and verify.
It is not a generic website, an AI guess, or an open-ended permission slip. The exact source and its scope matter.
If the source is missing, expired, conflicting, out of scope, or otherwise unresolved, CharterGate fails closed instead of inventing permission.
Choose the level of detail
The modes change presentation and workflow depth, not the governing authority or the meaning of the result.
For people who want plain-language guidance without needing to know the internal field names.
For recurring work where the same source, limits, conditions, and evidence need to stay understandable and consistent.
For operators who need deeper diagnostics, bindings, provenance, and audit information.
Templates and examples
Templates help structure common requests. They do not create authority and they do not pre-authorize an outcome.
Check the exact content, destination, owner approval, and boundaries before a separate publisher acts.
Evaluate the proposed action and authority before the agent hands work to an executor.
Check actor, target, scope, conditions, and current authority before a separate system changes data.
AI governance
It can govern an AI-controlled workflow without becoming the executor or the source of organizational authority.
The exact requested action is assembled with its target and scope.
Permission comes from registered authority outside the model being governed.
CharterGate resolves, compares, fails closed when needed, and records the result.
Only an independently authorized executor with its own credentials can perform the action.
Meet Gate Keeper
Gate Keeper helps people understand what CharterGate needs, explains conditions and evidence, and makes fail-closed outcomes easier to navigate. It is a presentation layer only.
Windows private preview
The private preview is focused on learning where governance evaluation adds clarity before consequential AI or workflow execution.