InterAIPre-execution authority
Action Boundary Lab

Propose an action
See whether it should execute

InterAI evaluates a proposed action before execution. This lab never sends the email, moves the money, deletes the record, or performs the external side effect.

AGENT PROPOSES → INTERAI EVALUATES → ALLOW / REVIEW_REQUIRED / BLOCK → EXECUTION LAYER ROUTES
Step 01

Define the proposed action

Step 02 · access

Free safe demo evaluates only the fixed read-only sandbox fixture. For custom scenarios, paste your own InterAI API key; it stays in this page's memory and is sent only to /verify.

Nothing has executed.

Step 03

Authority decision

No decision yet. Choose a scenario and evaluate it.

Execution stays outside InterAI

A verification API is not an execution boundary by itself. Your framework, orchestrator, wallet, workflow engine, or tool runner must make execution conditional on the returned authority decision. Hosted authenticated execution can also apply account policy outside the request body: HOST → ACCOUNT → CALLER → EFFECTIVE.