Trust & Compliance

Trust & Compliance

Measure honestly, report honestly — writing the definitions down is a business choice we made on purpose.

Data security

  • Strict tenant isolation: business data is isolated by tenant and project; cross-tenant access is refused at the architecture layer.
  • Zero public training: customer data is never used to train public models.
  • Encrypted in transit and at rest; sensitive credentials are stored encrypted (app-specific credentials only — we never store your platform login passwords).
  • Security baseline self-assessment benchmarked to MLPS Level 3: 46 check items, continuously self-reviewed (a self-assessment mechanism, not a certification).

Content compliance

  • Two-layer sensitive-word system: platform rules + tenant lexicons.
  • 23 vertical industry compliance packs (healthcare, finance, education rules built in).
  • AI-generated content labeled as required.
  • Advertising-law red-line checks: absolute claims and effect promises are intercepted before publishing.
  • Transparent definitions: every external metric carries a denominator and boundary; unsampled is never shown as 0.
Standards & regulations we follow

Compliance is built into the product, not attached afterwards. It is enforced in code against national and industry standards.

GB 45438-2025
Labeling method for AI-generated content (mandatory national standard): explicit plus machine-readable labels on AIGC content, with a separate ledger recording whether the label actually displays on each platform.
GB/T 45654-2025
Security requirements for generative AI services: implementation requirements for corpus safety, generated content and security assessment.
T/CAPT 026—2026
The GEO-specific association standard led by Xinhua Net (the online arm of China's state Xinhua News Agency): nine traceability elements for content; corpus poisoning and fake-consensus black-hat tactics are exactly what the platform blocks.
AIIA/T 0277—2026
the China Academy of Information and Communications Technology (CAICT) "Trustworthy GEO service requirements": 5 aspects × 10 requirements — the platform's governance, material review and traceability processes are built against it.
T/CGCC 119—2026
China General Chamber of Commerce GEO service specification: six prohibited technical behaviors and four core metrics, written into delivery and acceptance.

The platform runs compliance self-checks against these standards; certifications and ratings are published only once obtained — no unearned claims.

What we don't do

A vendor that writes down what it refuses to do is the only kind that puts legal and procurement at ease. Every line below maps to an actual constraint in the platform's code — not copywriting.

  • We never store your third-party platform login passwords — channels that need authorization ride revocable OAuth token custody.
  • We don't sell engagement-farming "action lists" — on UGC platforms we only monitor and correct assets you own.
  • We don't imply "add structured data and you will be cited" — no causal promises between schema and citations.
  • We don't make unverifiable comparisons — no "guaranteed rankings", no "industry No.1" claims that cannot be substantiated.
Effect boundaries

We commit to honest measurement and honest reporting — not to rankings, citation rates or business outcomes.

Credentials & certifications

Relevant credentials and certifications will be published here once obtained.

The platform is building its governance and processes against the China Academy of Information and Communications Technology (CAICT) "Trustworthy GEO service requirements" (AIIA/T 0277—2026), 5 aspects × 10 requirements; evaluation and certification credentials are published once obtained.