MeetHidden tests
The overselling case, the duplicate request, the failure path the brief never mentioned. They run against the submission in a clean environment and never reach the browser.
Hire engineers whothink ahead.
Every candidate gets the same challenge and the same AI agent. We grade the code they ship.
Built for the stack your team already ships.
Built for your preferred stack
Meet Qwack
Qwack is our deliberately imperfect coding agent. It’s lazy, passive, cuts corners, doesn’t care much about clean code or good design, and sometimes gets things wrong. You wouldn’t want to own the code it hands back. A strong engineer knows how to guide it, challenge it, and turn its output into something worth shipping. That’s what Ducker measures.
See how Qwack behavesAgentic code review
A review agent reads the finished implementation the way a senior reads a pull request — structure, error handling, data modelling, naming, and the side effects a careful engineer would have caught.
Candidate report
A score per dimension, what the submission handled, what it missed, and a hiring recommendation you can check against the diff.
Follow-up interview
Qwack asks why they made the calls they made, in their own code, while it is still fresh.
Build your own challenge
Describe the use case, pick the stack. An agent writes the brief, the starter code, the hidden tests and the rubric.
Building…
export const challenge = defineChallenge({
stack: 'nestjs + postgres',
brief: 'Add seat holds to the booking API',
hidden: [
'overselling under concurrency',
'retries stay idempotent',
],
});Describe another challenge
It lands where your team already works
Send from your ATS, get the score back on the candidate.
Explore integrations
Or build against the API
Create assessments with one authenticated call, receive a webhook on start, submission and grading, and pull the full report as JSON.
Plans & Pricing
Start with no subscription and pay only when a candidate starts an assessment. Move to Growth once hiring becomes routine.
No credit card to start. On Demand charges only for assessments a candidate actually starts — counted when the device check passes and the timer begins.
FAQ
The rubber duck talks back now.
Early access / now open
Meet the developers you'll want to build with.
Get early access
No credit card · Pay only for assessments candidates start