Sample candidate replay
See the evidence your reviewers get.
This sample mirrors Kagento's review surface for a completed candidate assessment: score, tests, duration, terminal replay, scorecard export, and internal reviewer notes. It is a demo replay, not an AI interviewer.
Bookstore API Repair
Backend assessment - completed
24m
- Score
- 92/100
- Passed
- 23/25
- Duration
- 24m 18s
- AI setup
- Candidate-owned tool
Evaluation complete
Bookstore API Repair
92/100
Passed
23/25
Failed
2/25
Duration
24m 18s
Try againLeaderboard
Session Replay
Terminal Replay
Complete shell input and output.
replay - 00:17:36
$ pytest -q
2 failed, 23 passed
$ sed -n '40,70p' api/routes/books.go
return json.NewEncoder(w).Encode(books)
$ git diff -- api/routes/books.go
+ if books == nil { books = []Book{} }
$ pytest tests/test_books.py::test_pagination -q
FAILED expected page=2 offset=20
$ vim api/routes/books.go
manual correction: offset = (page - 1) * limit
$ pytest -q
25 passed
Reviewer actions
View Session ReplayPrintable scorecardDownload JSONSubmitted May 6, 2026
Human-reviewed
Kagento records evidence. Your engineers decide.
No AI interviewer
The platform records candidate work; it does not conduct or grade interviews by itself.
Candidate-owned tools
Candidates use the AI setup they would use at work.
Deterministic tests
Scores come from task tests, not AI judgment.
Want to run this on your own role?
We can set up one task, invite flow, replay review, and scorecard export for a pilot.