Ansvar AI

About Ansvar AI

Ansvar AI provides compliance and security intelligence to AI assistants over the Model Context Protocol (MCP). A customer connects their existing AI client; Claude, Microsoft Copilot Studio, ChatGPT, or any MCP-compatible agent — to a single gateway (gateway.ansvar.eu) via OAuth 2.1. The gateway routes questions across a fleet of specialised servers covering national legislation, EU regulations, court decisions, preparatory works, agency guidance, and technical standards. Coverage spans 29+ live jurisdictions (27 European countries, the United Kingdom, and the United States), with more than 240,000 national laws and 4.2 million individual provisions, plus EU-level corpora for frameworks such as GDPR, NIS2, DORA, MiCA, PSD2, the EU AI Act, CRA, MDR, IVDR, the Machinery Regulation, UN R155/R156, CSRD, and EUDR. Live threat intelligence (CVE, KEV, EPSS) and a canonical control library (NIST 800-53, CSF 2.0, ISO 27001, NIS2, DORA, CRA crosswalks) support security and risk work. Three principles distinguish the design. First, grounding: every served result includes a paragraph-level citation with source URL, publisher, and licence; content is ingested only from original, authoritative publishers, and the live tier is 100% source-licensing audited. Second, accuracy over availability: if a data source is unreachable, the gateway fails with an explicit error instead of letting the assistant answer from model memory. Third, auditability: structured workflows (threat modelling with STRIDE/LINDDUN, TARA, DPIA, regulatory gap analysis, audit packages) produce citable deliverables, and uploaded customer documents are cited at paragraph level with content hashes. Five subscription tiers (Free, Solo, Premium, Team, Company) scale from single-jurisdiction research to organisation-wide use with document upload, pooled workflow allowances, and a per-tenant audit ledger. Per-standard add-ons cover licensed standards content.