Artificial Intelligence Evaluation for Guardrails, Integrity & Safety. A standardized, repeatable framework that turns AI safety from a marketing claim into a measurable, comparable, defensible rating.
Every layer of the A.E.G.I.S.™ platform is operational, from evaluation submission and validator consensus through to certified badge issuance, real-time dashboards, and full API access.
A full-featured React SPA at dashboard.rateai.ai. Track products, test runs, prompt cases, validator votes, pillar scores, certifications, and your audit ledger, all in one place with live charts powered by Recharts.
Full JWT and API-key authenticated REST API at api.rateai.ai. Register products, submit evaluations, query scores and certifications, access the public registry, and integrate into any CI/CD pipeline.
TOTP multi-factor authentication with backup codes. Google SSO with auto-account linking. Role-based team management with email invite flow: Owner, Admin, Analyst, and Viewer roles.
Every prompt case, validator vote, consensus decision, and certification is chained via SHA-256 hash. Records are write-once: nothing can be altered, deleted, or backdated. Full audit log via dashboard and API.
Pillar weights, A-Scale thresholds, and quorum parameters are never hardcoded; all are defined by the active scoring epoch. Epochs freeze scoring parameters for reproducibility, keeping evaluations comparable across time.
Every prompt case captures response time in milliseconds. The dashboard surfaces per-pillar average and p95 latency charts and response-time trends per product, giving visibility into both safety and performance.
Full subscription management with Stripe integration. Plans, billing events, and idempotent webhook handling built in. Manage your subscription, view invoices, and upgrade from the Settings page.
Anyone can verify a RateAI certification without logging in. Share your A-Scale badge publicly: customers, partners, and regulators can independently confirm the certification is current and has not been downgraded.
Full internal backoffice at admin.rateai.ai for RateAI staff: company management, user oversight, product and evaluation visibility, admin user CRUD, and scoped API-key management with super_admin / support / read_only roles.
Each pillar is scored independently through automated test batteries, adversarial simulation, and expert review, then combined into a single A-Scale rating.
Does the system state falsehoods as facts? We benchmark factual reliability under pressure, ambiguity, and domain-specific questioning.
Can the system be manipulated into unsafe behavior? Jailbreak attempts, prompt injection, and boundary-pushing scenarios test every guardrail.
What does the system collect, retain, and leak? We audit data collection behavior, PII exposure, and privacy compliance posture.
Does the system behave appropriately with vulnerable users and sensitive topics? Ethical boundary stress tests measure alignment in practice.
How far can the system act on its own, and can it be stopped? We evaluate autonomous behavior, escalation paths, and human override integrity.
How does the system hold up against a motivated attacker? Red-team exercises and multi-agent simulations probe real-world resilience.
Can the system's behavior be understood and audited? We assess traceability, disclosure quality, and the evidence trail behind decisions.
Pillar results determine the A.E.G.I.S.™ Certified seal a system is awarded, from A-5 Guardian, the highest safety assurance, down to A-1 Unrestricted.
A-5 · Guardian
A-4 · Shielded
A-3 · Contained
A-2 · Exposed
A-1 · UnrestrictedEvery certified system receives a standardized A.E.G.I.S.™ scorecard: a safety label as readable as a nutrition panel, built for procurement teams, regulators, retailers, and parents.
The headline is the Overall Safety Rating: a single letter grade from A (Maximum Safety) to F (Dangerous) that anyone can read at a glance, no technical background required.
Beneath it, every one of the seven A.E.G.I.S.™ pillars is graded individually (Hallucination Integrity (H), Guardrail Resistance (G), Data Handling & Privacy (D), Ethical Alignment (E), Autonomy & Control (C), Adversarial Robustness (R), and Explainability & Transparency (X)), so buyers see exactly where a system is strong and where caution is warranted.
Certified products carry this label alongside the “RateAI Safety Certified” badge and a verifiable listing in the public Safety Registry.
Consumer products are additionally certified for age appropriateness, from Infant to Adult, with supervision guidance printed directly on the mark.
We define the evaluation scope: agents covered, autonomy level, deployment context, and the external reliance the certification must support.
Standardized test batteries run across all seven pillars: guardrail probing, hallucination benchmarks, data collection audits, and ethical stress tests.
Red-team exercises, multi-agent simulations, and expert review panels validate automated findings for higher-assurance engagements.
The system receives its A-Scale rating, certification badge, formal approval letter, and public Safety Registry listing.
As models retrain and products evolve, continuous monitoring and periodic reassessment keep the certification current and defensible.
Start with a safety benchmark for one agent, or scope a full certification engagement.
Automated assistant.
A person is one step away.
Your chat is saved so our team can help.
Or email support@rateai.ai.
