|
Razon Lab | Experiments that earn their claims
Razon Lab is a public evidence ledger. See how AI, automation, and security tests are bounded before a public record is released.
|
79 % |
|
Sensyscope — Software, verified.
A bounded security review for AI-built applications: deterministic findings, founder-led review, source evidence, and explicit gaps. Not a pentest or safety...
|
78 % |
|
AlientAI — Researching Signal in a Noisy Market
AlientAI is a research-first market intelligence project building point-in-time datasets and testing selective stock-ranking models on unseen future data.
|
78 % |
|
Taylor Sutherland — AI Assistant Quality Audits
Taylor Sutherland provides independent, human-reviewed quality audits for customer-facing AI assistants. Explore the test method and discuss a focused pilot.
|
77 % |
|
Referee | A transparent reliability score for every scholarly paper
Referee is research-integrity infrastructure for high-volume scholarly publishing, turning technical evidence into transparent, machine-readable reliability...
|
77 % |
|
SoulDriver - Deliberative Intelligence Platform
KARP - Four frontier AI models research your question in structured deliberation, cross-examining every claim. Cited, board-ready reports you can trust.
|
77 % |
|
Perspective Cosmology | Perspective Cosmology
An experiment in AI-driven theory generation. Run #1 is archived; run #2 rebuilds on machine-checked proof.
|
77 % |
|
unerr | Assurance for AI work
Examine a defined AI task, record the evidence, and state what the result does and does not support.
|
77 % |
|
GoodHart Labs | The frontier for clinical-AI reliability
GoodHart Labs is the independent institute for validation, custom benchmarking, and verification of AI that touches patients. When a hospital asks whether yo...
|
77 % |
|
Test Case Generator AI by Adaps
AI-powered test case generation, powered by adaps
|
77 % |
|
Cognition Observatory
Documenting how AI systems fail in extended use, and what that means for the people relying on them.
|
77 % |
|
PackAudit — packaged commodity inspection
Inspection software that checks packaged-commodity labels against Legal Metrology rules.
|
77 % |
|
Invigil — the score a lab cannot train against
Invigil measures how well frontier AI models reason, on formal verification problems written in Lean 4 that the models have never seen. A proof checker decid...
|
77 % |
|
Prufing — the audit agent that shows its work
Prufing tests every control against your evidence, cites the exact line it relied on, and holds every verdict as a draft until you sign it.
|
76 % |
|
Zua.ai — Intelligence you can explain, audit, and trust
Zua.ai is a risk analytics platform that layers rules, statistics, machine learning and AI assistance into one governed intelligence layer — so every score,...
|
76 % |
|
BluelightAI - Interpretability for AI systems
BluelightAI reads your test and production traces at the concept level and surfaces the behavior your evals never described, so you can see why a model faile...
|
76 % |
|
Ownable — the seal on everything you ship
Ownable governs AI-built software from idea to release — every change gated, every rule registered, every release sealed with evidence.
|
76 % |
|
Kansaro, now a Workiva company — Testing automated. Judgment amplified.
Cut audit and control testing by 50% or more with AI automated evidence review, attribute testing, tickmarks, and suggested conclusions — now part of Workiva...
|
76 % |
|
EVALution — AI evaluation for teams shipping LLM features
EVALution helps companies integrating AI measure whether it actually works: evaluation frameworks, pre-launch readiness audits, and regression evals that run...
|
76 % |
|
RedTeamWorks | Offensive Security Consulting
RedTeamWorks provides offensive security consulting, penetration testing, attack-path validation, and practical remediation guidance.
|
76 % |
Detected social media networks