Plant a safe canary
Every application gets a harmless, unique instruction that only a page-reading agent would encounter.
Trust infrastructure for modern hiring
Trust the signal.
Question the agent.
Generative AI broke hiring on both sides. Trippy gives recruiters an evidence-based agent safety signal—and gives candidates a way to verify the people hiring them.
01 · Untrusted instruction02 · Agent action03 · Explainable signal
One narrow signal
Observe what happened. Don’t pretend to know why.
Validate the action
The server checks whether the session selected its exact nonce-bound decoy. No fuzzy AI detector involved.
Offer a fair next step
Recruiters see evidence. Candidates get a transparent verification path. Nobody is automatically rejected.
Both sides of the funnel
A trust layer—not another hiring platform.
Trust coming in.
See when a delegated browser agent follows an untrusted page instruction, then route that session to human review.
For recruitersTrust going out.
Compare a job page or recruiter domain against public company sources and surface the evidence a candidate should inspect.
For candidatesFairness is a feature
A signal for review. Never a verdict on a person.
No automatic rejection
No résumé AI detection
Evidence first
Remediation always