dVeracity

Check data against the standards, canonical models and policies it claims to conform to, and get back a graded, machine-readable result that says what was checked, what was not, and why.

dVeracity is a verification engine for regulated data. An agent can read a canonical data model at design time (entities, sectors, semantic codes and policy guardrails), validate a payload against a named standard before submitting it, look up approved regulatory compliance policies with their rules and provenance, and query verified datasets in natural language. Every result states its model version and the checks it ran, so it can be handed to a reviewer rather than pasted into a chat.

The first standard onboarded is Open Footprint (The Open Group), with O-DEF semantic codes, PACT product-footprint validation and verified emissions and methane datasets. The engine is domain-neutral; further standards are onboarded through the same path.

Eighteen tools, all read-only. Model, sector, semantic-code and standards-catalogue tools answer from a version-controlled snapshot of the standard. Natural-language queries, validation and compliance-policy lookups run against the caller's dVeracity account; when a tool cannot run for the current account it says so in a machine-readable way and tells the agent to ask its human operator, and never retries on its own.

Authentication is OAuth 2.1 with dynamic client registration; a human approves access on dVeracity's consent screen. Optional GLEIF-rooted vLEI verified-agent identity is supported for organisations that need attributable agents.

Built for compliance, ESG and reporting teams, standards implementers, and anyone whose data has to be checked before it is trusted.