Hire and assess developers for the age of AI agents.
The same tests to hire candidates and to measure your team: from the job description, with real projects, AI allowed and assessed, and a report with each person's score.
- For candidates and for your team
- AI allowed and assessed
- Nothing to install, in English or Spanish
María Rodríguez
Problem 1 of 2 · Total by category
Cases
- Empty listnot run
- One categorynot run
- Several categoriesnot run
Activity
Tests in 38 languages and 29 frameworks
More than 450 ready-made tests, including 117 exams on AI, agents and automation.
- +43
Assess the work the way it's done today
The tests mirror the daily work of a team that already works with AI.
- 01
Real work
Multi-file projects with a terminal and a dev server, graded with automated tests in a clean environment.
- 02
AI fluency
In project tests, the candidate works with an assistant and you see how they used it: what they asked, what they reviewed, what they verified.
- 03
Integrity
Pastes, tab switches and typing rhythm, recorded with rules. No camera and no recording.
- 04
Explainable results
A score per case against your threshold and an AI conclusion that helps you decide, without changing the score.
From the job description to the decision, in one place
Four steps your team repeats in every hire. Everything stays saved in the interview.
Job description
We're looking for a Frontend Engineer for our product team. Responsibilities · Build interfaces in React + TypeScript · Consume REST APIs and handle async state · Care about performance and accessibility Requirements · Strong in modern JavaScript · Experience with hooks and testing · Comfortable with Git and teamwork
Frontend — TypeScript
Code test · Mid · not generated
Pick a level and generate the test to see the problems.
It's saved as a draft. You review the problems, the cases and the passing score, and publish it when it's ready.
Illustrative example: in the app, generating and checking the test takes a few minutes.
A format for every position
Each position uses one test. They all grade themselves on submission.
Code
- How it's graded
- Input and output cases, or cases on a function; some visible and some hidden.
- What the candidate uses
- Editor that runs the code in the browser.
- Technologies
+32
38 options
Exam
- How it's graded
- Multiple choice, with topic and difficulty.
- What the candidate uses
- One question at a time, against the clock.
- Technologies
- 117 exams on AI and current tools
Project
- How it's graded
- Hidden tests that run on submission.
- What the candidate uses
- Several files with an explorer. AI allowed if you turn it on.
- Technologies
+1
7 options
Project with frameworks
- How it's graded
- Automated tests in a clean environment.
- What the candidate uses
- Dev server, terminal and live preview.
- Technologies
+23
29 options
Build your team for the age of AI
Hire with tests that mirror today's work and measure your people with the same tests.
Companies
Hire by area, with your brand
Organize hiring in projects. Each position has its test, and recruiters and interviewers only see their own work.
- Your subdomain, logo and color
- Five roles with clear permissions
- Ranking by project and CSV export
Agencies
One project per client
Each client is a project with its positions and its assigned team. If a client needs its own brand, it goes in a separate company and you switch without signing out.
- Projects and positions per client
- Recruiters assigned by project
- Several companies in one account
Team assessments
Measure your team with the same tests
Combine tests with the weights you want, share a link restricted to your company's email domain and build level programs.
- Assessments with several tests
- Level programs with a certificate
- AI adoption survey by team
Clear signals, without watching anyone
What happened in the session is recorded with rules, not AI. They're points to talk through with the candidate, not an accusation.
- Code runs isolated
- Every run goes to an isolated environment with no network. Framework projects run in a virtual machine that's destroyed at the end.
- No camera or recording
- We record pastes, tab switches and typing rhythm. No video and no screen capture.
- One link, one device
- The test is tied to the browser where it starts. To continue on another machine, the candidate confirms with a code sent to their email.
- Your space, your data
- Each company works on its own subdomain, separate from the rest, and each role only sees what it needs.
For your team
Recruiters organize, interviewers watch and decide, and leadership sees the ranking of each project.
- Tests from the job description or the catalog
- Live view of the editor
- Printable report with a conclusion
For the candidate
A fair, frictionless experience: they open a link and work the way they do every day.
- No account and nothing to install
- In English or Spanish
- Editor, terminal and live preview in the browser
Simple pricing, no surprises
A fixed plan, monthly or annual. Every plan has every feature; anything over what's included is billed per unit.
Starter
To start assessing with technical tests.
USD 79per month
Billed every month
Create accountIncluded per month
- 10 code interviews
- 10 assessment participants with code
- 500 exams
Over that: USD 8 per interview or participant · USD 0.25 per exam
AI and authoring per month
- 1,000 AI conclusions
- 20 AI-generated tests (5 with frameworks)
- 1,500 validation submissions
Team
RecommendedFor teams that hire and assess their people every month.
USD 299per month
Billed every month
Create accountIncluded per month
- 50 code interviews
- 50 assessment participants with code
- Unlimited exams
Over that: USD 6 per interview or participant
AI and authoring per month
- 3,000 AI conclusions
- 80 AI-generated tests (20 with frameworks)
- 6,000 validation submissions
Scale
For high volume and several clients or areas.
USD 699per month
Billed every month
Create accountIncluded per month
- 150 code interviews
- 150 assessment participants with code
- Unlimited exams
Over that: USD 4 per interview or participant
AI and authoring per month
- 6,000 AI conclusions
- 200 AI-generated tests (50 with frameworks)
- 15,000 validation submissions
How it counts
1It counts when it starts
An interview or a participant counts when the person starts. Scheduled or cancelled ones don't count, and each participant counts once per assessment even if they retake (in a program, each level is an assessment).
2Going over never stops you
Extra interviews, participants and exams are billed as extras when the period closes. You never leave a candidate waiting.
3AI pauses at the cap
Conclusions, generations and validations are never billed as extras: when used up, they pause until the next period.
Prices in US dollars, taxes excluded. Need more than Scale includes, or several companies? Let's talk.
Frequently asked questions
Hire and assess by the same standard.
Candidates and team, with tests that mirror how work gets done today. Create the account, choose the plan and build your first test.