Semantic API
Verified emissions data,
built for AI agents
Ask questions in natural language against a knowledge graph of verified emissions data — Open Footprint, PPDM and OGMP 2.0 methane observations — and get answers your agent can act on. Metered per call, with payment semantics agents actually understand.
What your agent can ask
- “Which countries had the most detected methane plumes last year, and which operators do they attribute to?”
- “Total verified Scope 1 emissions for this facility, by period.”
- “Validate this payload against the Open Footprint data model and tell me what fails.”
Behind the API: virtual knowledge graphs over the Open Footprint canonical model, PPDM, and UNEP IMEO methane data — 24,900+ verified detection events and thousands of sites, companies and reports — plus Validation-as-a-Service against sustainability standards.
Three ways in
MCP server
One command for Claude Code, Cursor, or any MCP client. Tools for querying, validation, and credit balance.
Plain HTTPS
An API key and a POST. The manifest documents every endpoint, cost and error contract, machine-readably.
Verified identity
Optional vLEI agent credentials: every call attributed to a GLEIF-rooted legal entity.
Quick start (MCP)
# 1. Subscribe (api tier) at https://dveracity.com/pricing, create a key (dvrc_...)
# 2. Add the server:
claude mcp add dveracity -e DVERACITY_API_KEY=dvrc_yourkey -- npx -y @dveracity/semantic-mcp
# 3. Ask your agent about emissions. Out of credits? The API answers HTTP 402
# with a purchase pointer — your agent relays it instead of failing weirdly.Agents discover the full contract at /api/v1/semantic/manifest and a plain-text overview at /llms.txt.
Honest billing for autonomous callers
Metered calls cost credits. When an account runs out, the API returns HTTP 402 Payment Required with a machine-readable body: balance, credits needed, and the purchase endpoint. Well-behaved agents surface that to their operator and stop — no retry loops, no mystery failures.
{ "error": "insufficient_credits", "balance": 0, "needed": 1,
"purchaseUrl": "/api/v1/vaas/credits/purchase" }Verified agents, verified data
The differentiating layer: an organization can issue its AI agent a dVeracity Agent Authorization credential— an ACDC chained to the organization's vLEI, rooted at GLEIF. The agent authenticates by verifiable presentation, and every metered call is attributed to the agent's identifier and the legal entity's LEI in the audit trail. Your auditors can verify not just the data, but every agent that touched it. API keys keep carrying billing; the credential carries identity and scoped authorization, revocable by the issuing organization at any time.
Standards inside
Built on the Open Footprint® data model with O-DEF semantic annotations, PPDM for upstream assets, PACT for product footprints, and OGMP 2.0 / UNEP IMEO for methane — with verification workflows aligned to ISO 14064-3. Relevant if you are building for CSRD, CBAM, CFR or TIER compliance.