Hackathon judging you can defend.
Juryza runs your whole hackathon — registration, teams, submissions, judging, community voting and published results — and makes the ranking statistically fair. A harsh judge can't sink a project, and every score is on the record.
juryza.local/e/sample-hack-2026/leaderboard
Judged leaderboard
41 projects · 30 judges · 123 reviews
- 1
Issue Radar
Night Owls
+1.62▲3raw 3.70
- 2
Carbon Ledger
Greenfield
+1.31—raw 4.05
- 3
Loop Assist
Pair of Docs
+0.94▲2raw 3.45
- 4
Quiet Hours
Deep Work
+0.52▼3raw 4.15
Impact
30%
Execution
35%
Innovation
20%
Demo
15%
Live on this instance
- Events hosted
- 0
- Projects submitted
- 0
- Participants
- 0
- Judges on panels
- 0
- Scores recorded
- 0
- Community votes
- 0
Everything in the box
From first sign-up to signed certificate
One self-hosted app covers the whole event. No plugins to buy, no spreadsheet glue, no per-seat pricing.
Events & tracks
Dates, tracks, prizes and rules in one place. The lifecycle follows the calendar — no manual phase flipping.
Teams by invite link
Create a team, share one link. Max team size and deadlines are enforced by the backend, not the button.
Notion-style write-ups
A block editor with slash commands for project pages — headings, checklists, code, embeds. Drafts until submit.
Weighted rubric
Organizer-defined criteria and weights. Judges score what matters; weights normalize to 100% automatically.
Backend role isolation
A judge can't read another judge's scores — not in the UI, not with a raw curl. Enforced on every request.
Z-score normalization
Each judge's scores are standardized against their own mean and spread, so harsh and lenient judges count equally.
Pairwise Bradley–Terry
Prefer comparisons to absolute scores? Judges pick A vs B and a Bradley–Terry model turns that into a ranking.
Quadratic community voting
Voters spread a credit budget; n votes cost n². Enthusiasm counts, but a single fan club can't buy the prize.
Similarity & comparison
Duplicate and near-duplicate submissions are flagged automatically, with side-by-side project comparison.
Readable audit trail
Every submission, score, publish and permission change is logged with actor and time — in plain language.
REST API, webhooks, OpenAPI
Every UI action is an API call. Bearer tokens, signed webhooks and an OpenAPI 3.1 spec for your own tooling.
Verifiable certificates
Signed certificates for winners and judges. Anyone can check a serial — tampered records fail verification.
Embeddable gallery
Drop the public project gallery into your event site. Search and track filters come with it.
CSV / JSON import-export
Bulk-import projects, export scores and results. Your data leaves in open formats whenever you want.
Built to be read by machines, too
Browse the full API reference, or pull the spec straight into your client generator.
How it works
Five steps, one timeline
Every event moves through the same lifecycle. Phases follow your dates, so nobody forgets to close submissions at midnight.
- 1
Create
Set dates, tracks, prizes and a weighted rubric. Publish when ready.
- 2
Submit
Teams form by link and write their project up. The deadline is enforced server-side.
- 3
Judge
Assignments are balanced across the panel. Judges score blind to each other.
- 4
Vote
The community spends quadratic credits on a randomized ballot.
- 5
Publish
Normalized results stay hidden until you publish. Certificates are signed.
Judging integrity
Luck of the draw shouldn't decide the winner
With 40 projects and 30 judges, no one sees everything. If your project lands with the toughest judge, a raw average quietly punishes you. Juryza compares each score to that judge's own habits first.
Per-judge standardization
z = (x − μjudge) / σjudge
Then the weighted rubric is applied and z-scores are averaged per project. Ties and single-review projects are handled explicitly — the method is documented, with a worked proof.
- Judges only ever see their own assignments and scores — enforced by the API.
- Pairwise Bradley–Terry mode for events that prefer comparisons to scales.
- Every score change lands in the audit trail, with who and when.
Happening here
Events on this instance
FAQ
Questions organizers ask
Something else? The API reference documents every rule the backend enforces.
Your next hackathon, judged fairly — running in one command.
Self-host it on a laptop for a weekend event or a server for a season. Seeded with sample data so you can click through everything first.