LQS Enterprise · for Fortune 500 procurement

The quality standard your risk team will cite by name.

LabelSets LQS is a 19-dimension training-data rating with cryptographic signature — engineered for SR 11-7, EU AI Act Art. 10, §1557, and FDA 21 CFR 11 paperwork. Run it on our marketplace data, your internal data, or both. Cite the cert hash. Defend the model.

Talk to enterprise → Read the methodology →
Cert fields map to the compliance paperwork your risk team already files
SOC 2
Type II · audit
HIPAA
BAA ready
EU AI Act
Art. 10 aligned
SR 11-7
Fed · model risk
21 CFR 11
FDA · e-records
GDPR · DPA
Art. 28 + SCCs
What you get

Three deliverables your team can cite directly.

Not a slide deck. Not a "platform." Three concrete artifacts your model risk, compliance, and procurement teams pin to their files.

①
A signed cert per dataset
19-dimension LQS score, 95% confidence intervals, contamination check vs 40+ benchmarks, oracle agreement (Fleiss κ). Ed25519-signed. Verifiable offline by anyone with the public key — no LabelSets in your trust chain.
cert_hash · composite_lqs · signature_b64
②
Private-mode scoring (your VPC)
Score your internal training data against the same LQS the marketplace uses. Runs in your AWS / GCP / Azure tenancy. Data never leaves your perimeter; we ship a Docker image and a CLI. Your security team's favorite vendor surface area: zero.
Docker · CLI · air-gap option · BYO-key signing
③
Methodology paper + audit support
Citable methodology PDF for your SR 11-7 / EU AI Act / §1557 / 21 CFR 11 paperwork. We answer auditor questions directly under our DPA. Custom calibration corpora available for regulated domains (finance, clinical, legal).
PDF · DPA · auditor support · custom calibration
Compliance

The checklist your vendor-review team will hand you.

Pre-built for the paperwork your procurement, security, and legal teams need before they sign off.

SOC 2 Type II
In progress, target Q3 2026. Control mapping documented; attestation letter shared on request once issued. Until then, see /security for the live OWASP control posture.
status · in progress
HIPAA BAA
Business Associate Agreement scoped during enterprise engagement for clinical-tier datasets. De-identified data + optional subgroup masking.
coverage · clinical tier
Dedicated DPA
GDPR + CCPA Data Processing Agreement scoped during enterprise engagement. SCC clauses + sub-processor list provided on signature.
EU · Art. 28 GDPR
SSO & team access
Google OAuth + email auth live today. SAML SSO + SCIM (Okta, Azure AD, Google Workspace) on the enterprise roadmap, target Q3 2026.
status · OAuth live · SAML Q3
EU AI Act alignment
LQS cert designed to drop into Article 10 data-governance documentation. Per-dim CI + subgroup equity tracking included in every cert.
article · Art. 10
Private-mode scorer
In-VPC deployment of the scorer is on the enterprise roadmap; first design partners onboarding through the pilot program. Public-key verification works offline today.
status · pilot · roadmap
Audit log API
Per-license and per-verify audit trail exposed via JSON API. CSV export today; Splunk / Datadog / S3 connectors on the enterprise roadmap.
retention · 7 years
Uptime SLA
Custom uptime SLA negotiated on enterprise contract. Live status page with historical data at /status; current trailing-30-day uptime visible there.
sla · contract-defined
Direct support
Email support with 1-business-day response today. Named CSM + Slack Connect channel + quarterly roadmap reviews available on enterprise contract.
response · 1 business day
Who buys this

The teams that actually need the cert.

Model risk
Banks + asset managers
SR 11-7 and OCC 2011-12 compliance officers citing dataset quality in model validation packages. Independent third-party attestation closes the "training data lineage" gap.
Typical use: Fraud detection models · credit risk scoring · AML classification · ECOA fair-lending audits.
Clinical AI
Health systems + device OEMs
HHS §1557 non-discrimination obligations + FDA 21 CFR Part 11 electronic-record integrity. The cert's subgroup_equity + signed provenance satisfy both.
Typical use: Radiology triage · clinical decision support · diagnosis chain-of-reasoning evals.
LLM labs
Foundation-model teams
Pre-training and fine-tuning teams needing benchmark-clean corpora. contamination_clean flags overlap with 40+ public evals, per-benchmark.
Typical use: Instruction tuning · RLHF preferences · safety-eval clean training data · retrieval corpora.
Public sector
Regulators + agencies
EU AI Act Art. 10 compliance officers. NIST AI RMF MEASURE 2.2/2.3 evidence. Procurement teams onboarding AI vendors.
Typical use: Vendor onboarding · AI audit evidence packages · Art. 10 data-governance filings.
Talk to enterprise

Email us. We reply in writing.

We don't run scheduled walkthroughs — most procurement, methodology, and compliance questions are better answered in writing your audit team can quote. The full pipeline is self-serve at /demo; if you need contract terms, scoping, or pricing for >$50k commitments, drop a note below and we reply within 1 business day.

We respond within 1 business day. We never resell your info.