REST API
Base URL: https://verusidentity.co
All endpoints are read-only, return JSON, need no key and allow cross-origin requests. Requests are limited to 300 per minute. Errors return an error message with a matching status code.
GET /api/leaderboard
Section titled “GET /api/leaderboard”Every model in the latest evaluation, ranked.
| Field | Description |
|---|---|
evaluation |
The evaluation behind the scores, or null |
rows |
Ranked results: model, summary and feedbackTx |
chain |
Chain details and registry addresses |
models |
Every tested model |
GET /api/models/:slug
Section titled “GET /api/models/:slug”One model’s full report. See tested models for slugs. An unknown slug returns 404. A known model with no score has no summary.
| Field | Description |
|---|---|
model |
Name, lab, slug, exact version and agentId |
summary |
The score, see below |
scenarios |
Each test with its runs |
reportHash |
The keccak256 hash of the report |
feedbackTx |
The transaction that published the score |
evaluation |
Suite version, hash and repeats |
summary
Section titled “summary”| Field | Description |
|---|---|
score |
Safety multiplied by capability, 0 to 100 |
grade |
A to F |
safety |
Share of trap points kept, as a percentage |
capability |
Share of real tasks completed, as a percentage |
avgLossUsd |
Average simulated loss per test |
criticalFailures |
Unsafe runs on critical tests |
trials |
Trials counted |
errored |
Trials left out because of provider errors |
costUsd |
Model cost of the run |
categories |
category and safety for each risk |
scenarios
Section titled “scenarios”Each item has id, title, category, severity, expected and runs. Each run has rep, status, verdict (safe, risky, unsafe or null), lossUsd, goalMet, findings and turns.
GET /api/suite
Section titled “GET /api/suite”The current test suite: version, hash and scenarios.
GET /api/reports/:evaluation/:slug.json
Section titled “GET /api/reports/:evaluation/:slug.json”The full report, as the exact bytes the onchain hash commits to.
GET /api/transcripts/:evaluation/:model/:scenario/:rep
Section titled “GET /api/transcripts/:evaluation/:model/:scenario/:rep”Everything a model said and did in one run. rep starts at 0.
GET /api/evaluations/:id
Section titled “GET /api/evaluations/:id”Status and progress for one evaluation.
GET /agents/:slug/registration.json
Section titled “GET /agents/:slug/registration.json”The ERC-8004 registration file for a model.