Electricity & power
ERC electricity Ft tariff by customer class — parsed from the official PDF, page & bbox kept.
th.energy.electricity.ft.generalReference data for energy, money and carbon — every value resolves to the exact document it came from. And product-compliance verification — every verdict cites the real standard, or it isn't made.
REST /v1/*remote MCPSDK · TS + PyPOST /v1/verifyGET /v1/ft?period=2025-01&customer_class=generalCite or refuse — every number arrives with its receipt (value, unit, effective date, source document), and an answer that can't resolve a citation returns no number at all.
One pipeline, one trust envelope — here's the shape of it. Every figure is a capability of the platform, not a usage stat.
What data is in here
Nine domains across five jurisdictions, all on one pipeline — and every series, whatever the domain, carries the identical cited, bitemporal envelope. Each card is a live namespace you can query by series_id.
ERC electricity Ft tariff by customer class — parsed from the official PDF, page & bbox kept.
th.energy.electricity.ft.generalEPPO daily retail price structure, component-by-component and sum-reconciled, plus LPG, NGV and MY fuel.
th.oil.retail.structure.gasohol_95.retailBank of Thailand reference buying / mid / selling per currency, plus SGD and MYR.
th.macro.fx.usd.midPolicy rate, CPI and GDP, plus MLR / MOR / MRR bank rates and government-bond yields.
th.macro.indicator.policy_rateRegional equity indices — SET, SET50 and SET100, plus Malaysia's KLCI and Singapore's STI.
th.market.equity.setThai gold reference — bar and ornament, buy and sell quotes.
th.commodity.gold.bar_sellThe energy→carbon bridge — TGO grid & fuel emission factors, IEA, EU CBAM defaults and real DEFRA factors.
th.carbon.grid.emission_factorA weighted effective-energy-cost index and next-period OLS trend forecasts with a confidence band — labelled and carrying their input observation ids, never fabricated.
th.derived.energy.effective_cost_indexHow a number earns its place
A tariff PDF, an oil-price table, a central-bank JSON feed. The raw artifact is archived content-addressed to object storage before a single value is parsed — so the original is always recoverable, byte-for-byte.
Range checks, period-over-period deltas, cross-source reconciliation and sum-checks gate every value. A disputed or low-confidence reading is held back, never served. The bias is deliberate — trust over coverage.
Every revision is appended along two time axes — when a value took effect, and when it was reported. Add as_of to any read and reconstruct exactly what the record said on that date.
Responses carry a human-readable citation and a source_doc_uri that opens the exact archived page. An AI agent grounded in Veracle quotes the citation verbatim — it cannot state a number it cannot trace.
What's the electricity Ft tariff for the general class right now?
The current ERC electricity Ft for the general customer class is 0.3672 THB/kWh, effective Jan–Apr 2025.
r2://veracle-raw-artifacts/erc/ft/8a1f….pdf
A citation-grounded answer. The model can only cite retrieved observations — it cannot invent a number.
The second product
Send a product to POST /v1/verify and get a verdict — block, warn, allow or indeterminate. Deterministic checks plus a grounded LLM mapper, closed-world cited: every finding resolves to a real standard, or it isn't made.
POST /v1/verify"product": { "title": "legacy adapter", "claimed_cert_ids": ["TIS 2540-2554"]}
Block — a check failed; do not ship without resolving it.
tisi.cert-withdrawnClaimed certificate TIS 2540-2554 was withdrawn on 2012-12-08.
TIS 2540-2554source: data.go.th (TISI open data)Upload a certificate image — OCR reads the cert codes, then the same cited checks run.
every series tracked against its cadence budget
BOT reference, last 16 weeks
every value carries this
"series_id": "th.energy.electricity.ft.general","value": 0.3672, "unit": "THB/kWh","effective_date": "2025-01-01","is_current": true, "source_id": "erc","source_doc_uri": "r2://veracle-raw-artifacts/erc/ft/8a1f….pdf"
ERC electricity Ft — general class, Jan–Apr 2025: 0.3672 THB/kWh. Source: ERC announcement (p.2).
source_doc_uri resolves to the exact archived PDF
remote server over OAuth — point any AI client at it
"veracle": { "transport": "http", "url": "https://mcp.veracle…/mcp"}
get_ft_tariff(period, customer_class)product compliance, cite-or-refuse — every finding cites a real standard
POST /v1/verify → block | warn | allowth.tisi · nbtc · eu.cbam →a citation-grounded LLM that cannot fabricate a number
one pipeline, many verticals
TS + Python — validated responses, full provenance
import { VeracleClient } from "@veracle/sdk";const v = new VeracleClient({ apiKey: "vk_…" });const ft = await v.getFtTariffs({ period: "2025-01" });
The same cited, bitemporal data, reached the way that fits your job — every surface returns identical data and citations; only the way you reach it differs.
Point any AI client at the remote MCP server, or call Ask Veracle in EN/TH. The model can only cite observations it actually retrieved — it cannot state a number it can't trace.
MCP + Ask VeracleBuild on the typed REST API and TS/Python SDKs for validated responses, or pull whole datasets as bulk Parquet snapshots on R2 — DuckDB-queryable, provenance intact.
REST + SDK + exportCross the energy→carbon bridge with cited emission factors, IEA references and EU CBAM defaults — every factor traces back to a named publisher, so a footprint defends itself in an audit.
Carbon estimate APIAdd as_of to any read and reconstruct exactly what was known on a past date, watch a per-series freshness SLI against its cadence budget, and walk any value's full lineage back to source.
Bitemporal + lineageVerify a product against TISI/มอก., NBTC and EU CBAM before you list it — every finding cites the real standard it was decided from, so a block or a warning is always explainable.
POST /v1/verifyStart anonymous — no key, no account. Add a key when you're ready; every tier returns the same data and the same citations.
Some datasets are real — DEFRA carbon factors, ECB FX, ERC electricity Ft, and the TISI + NBTC standards corpus — alongside synthetic fixtures. Citations, lineage and provenance are fully real throughout.