Data sources
Every fact on enology.ai names its source. This page names the sources.
Machine-readable version: /api/sources.
Registries
| Source | Publisher | What we take | License | Cadence | Last read | Records | Confidence | Status | Notes |
|---|---|---|---|---|---|---|---|---|---|
| eAmbrosia — EU register of geographical indications (wine) | European Commission, DG AGRI | Every registered PDO/PGI wine name with type, country, file number, registration date and product-specification link; the legal spine for EU appellations. | Commission reuse policy (Decision 2011/833/EU, attribution) | monthly | 82,847 | high | fresh | — | |
| EU PDO/PGI list (manual extract) | European Commission | Manual download of the EU geographical indications list; superseded by the eAmbrosia adapter (Group 2). | EU reuse policy (attribution) | yearly | not yet | — | high | off | Off — manual download source; no automated ingest |
| TTB COLA label images | US Alcohol and Tobacco Tax and Trade Bureau | The approved label images attached to each e-filed COLA (front, back, neck, strip). | Public domain (US federal record) | weekly | 1,914,410 | high | fresh | Label images synced continuously; served at labels.enology.ai | |
| TTB COLA Public Registry (filings) | US Alcohol and Tobacco Tax and Trade Bureau | Every federal Certificate of Label Approval: brand, class, origin, permit, structured label fields. | Public domain (US federal record) | weekly | 2,170,776 | high | fresh | Weekly discovery of new filings, continuously synced |
off: not currently ingested — see its note for why. pending: never read, and still within one cadence of being enabled. fresh: read within the last cadence window. due: last read between one and two cadence windows ago. stale: never read and more than one cadence past being enabled, or last read more than two cadence windows ago.
Derived from registries
| Source | Publisher | What we take | License | Cadence | Last read | Records | Confidence | Status | Notes |
|---|---|---|---|---|---|---|---|---|---|
| Decanter World Wine Awards (DWWA) results | Future plc (Decanter) | Decanter World Wine Awards (DWWA) medal results per wine per year: medal/score, wine name, producer, vintage, country, region, category. Not ingested. | Future plc / Decanter terms prohibit text and data mining and scraping | yearly | not yet | — | medium | off | Off — no ingest — Future plc terms prohibit text/data mining and scraping (2026-09-01) |
| GI rulebook rule extraction | enology.ai (derived from MASAF/INAO rulebook PDFs) | Structured production rules (permitted varieties, max yield, min alcohol, min aging, zone communes, styles, sugar limits) read from MASAF/INAO appellation rulebook PDFs with quoted evidence. | Derived work; underlying rulebook PDFs are public government documents | quarterly | — | high | fresh | gpt-5-nano bulk / gpt-5.4-mini escalation; evidence-or-reject. | |
| International Wine & Spirit Competition (IWSC) results | IWSC (International Wine & Spirit Competition) | International Wine & Spirit Competition (IWSC) medal results per wine per year: medal/score, wine name, producer, vintage, country, region, category. Awaiting permission. | robots permissive (ClaudeBot crawl-delay 3) but ToS unverifiable by bot | yearly | not yet | — | medium | off | Off — awaiting permission (G10) — robots permissive (ClaudeBot crawl-delay 3) but ToS unverifiable by bot; letter drafted; founder browser-check of ToS requested |
| International Wine Challenge (IWC) results | William Reed (International Wine Challenge) | International Wine Challenge (IWC) medal results per wine per year: medal/score, wine name, producer, vintage, country, region, category. Awaiting permission. | ToS requires prior written permission for commercial reuse | yearly | not yet | — | medium | off | Off — awaiting permission (G10) — ToS requires prior written permission for commercial reuse; letter drafted |
| Label barcodes (GTIN/UPC/EAN) | enology.ai (derived from TTB label images) | Barcodes decoded from label images; checksum-valid EAN-13/UPC-A become GTIN identifiers. | Derived work; underlying images public domain | daily | 737,375 | high | fresh | Automated barcode/GTIN decoding from label images; no AI model involved | |
| Label transcription and extraction | enology.ai (derived from TTB label images) | Verbatim transcript plus vintage, ABV, appellation and varietals read from each label image with quoted evidence. | Derived work; underlying images public domain | monthly | 628,028 | medium | fresh | AI-assisted label reading; runs on a manual cadence | |
| Mundus Vini results | Meininger Verlag (Mundus Vini) | Mundus Vini medal results per wine per year: medal/score, wine name, producer, vintage, country, region, category. Not ingested. | meininger.de robots.txt disallows ClaudeBot entirely + Content-Signal ai-train=no | yearly | not yet | — | medium | off | Off — no ingest — meininger.de robots.txt disallows ClaudeBot entirely + Content-Signal ai-train=no (2026-09-01) |
| No-op (wiring test) | enology.ai | Exercises the source-runner job and scheduler wiring; writes nothing. | n/a | yearly | — | low | off | Off — internal smoke-test source (scheduler liveness check) |
off: not currently ingested — see its note for why. pending: never read, and still within one cadence of being enabled. fresh: read within the last cadence window. due: last read between one and two cadence windows ago. stale: never read and more than one cadence past being enabled, or last read more than two cadence windows ago.
Statistical series
| Source | Publisher | What we take | License | Cadence | Last read | Records | Confidence | Status | Notes |
|---|---|---|---|---|---|---|---|---|---|
| FAOSTAT crops and trade | Food and Agriculture Organization of the UN | Grape and wine production, area and trade by country and year. | CC BY-NC-SA 3.0 IGO | quarterly | not yet | — | high | off | Off — awaiting scheduler (planned) |
| OIV world vitiviniculture statistics | Organisation Internationale de la Vigne et du Vin (OIV) | Production, area, consumption and trade statistics by country/year, alongside FAOSTAT in industry_stats. | OIV public "State of the vine and wine sector" annual report + public data portal exports | yearly | not yet | — | high | off | Off — manual workflow: annual report PDF tables + founder-placed CSV exports; dashboards are JS-only with no export endpoint; data request drafted to data@oiv.int |
| TTB statistical releases | US Alcohol and Tobacco Tax and Trade Bureau | Wine production, removals and inventories by month. | Public domain | monthly | not yet | — | high | off | Off — awaiting scheduler (planned) |
off: not currently ingested — see its note for why. pending: never read, and still within one cadence of being enabled. fresh: read within the last cadence window. due: last read between one and two cadence windows ago. stale: never read and more than one cadence past being enabled, or last read more than two cadence windows ago.
Catalog
| Source | Publisher | What we take | License | Cadence | Last read | Records | Confidence | Status | Notes |
|---|---|---|---|---|---|---|---|---|---|
| Alko product/price list | Alko Oy (Finnish Alcohol Retailing Monopoly) | Retail product catalog: name, producer, country, region, grapes, vintage, ABV, volume, price for the Finnish off-premise monopoly assortment, published as a price-list XLSX. | Alko open data (public price list) | monthly | not yet | — | medium | off | Off — price-list URL verification pending |
| LCBO product catalog | Liquor Control Board of Ontario (LCBO) | Ontario retail catalog: product name, producer, country, region, vintage, price. Not ingested. | LCBO site terms prohibit automated collection | weekly | not yet | — | medium | off | Off — ToS prohibits automated collection — will not ingest |
| Open Food Facts (wine subset) | Open Food Facts (community/nonprofit) | Crowdsourced product database wine subset: GTIN barcode, product name, brand, quantity, image URL. Feeds identifier_mappings GTIN corroboration against our own barcode reads; no facts promoted to canonical. | ODbL + DbCL (attribution; we store GTIN+name in staging and publish only our own derived match) | weekly | — | low | fresh | GTIN corroboration against our own barcode reads; throughput proven on Cloud Run 2026-09-04; identifier_mappings writes >10K rows require founder approval (G16) at run time | |
| SAQ product catalog | Société des alcools du Québec (SAQ) | Québec retail catalog: product name, producer, country, region, vintage, price. | SAQ site terms under review | weekly | not yet | — | medium | off | Off — terms review pending |
| Systembolaget product catalog | Systembolaget AB (Swedish Alcohol Retailing Monopoly) | Retail product catalog: name, producer, country, region, grapes, vintage, ABV, volume, price and image for the Swedish off-premise monopoly assortment. | Systembolaget open API terms (attribution; mission-compatibility clause — no promotion of consumption) | weekly | not yet | — | medium | off | Off — product API discontinued by Systembolaget (store-info API only; portal host dead as of 2026-09-01); no ingest |
| Vinmonopolet product catalog | Vinmonopolet AS (Norwegian Wine and Spirits Monopoly) | Retail product catalog: name, producer, country, region, grapes, vintage, ABV, volume, price and image for the Norwegian off-premise monopoly assortment. | Vinmonopolet open API terms (attribution; no resale of raw feed) | weekly | not yet | — | medium | off | Off — key registered 2026-09-01; adapter endpoint reconciliation pending (VERIFY-AT-KEY) before enable |
off: not currently ingested — see its note for why. pending: never read, and still within one cadence of being enabled. fresh: read within the last cadence window. due: last read between one and two cadence windows ago. stale: never read and more than one cadence past being enabled, or last read more than two cadence windows ago.
Export registry
| Source | Publisher | What we take | License | Cadence | Last read | Records | Confidence | Status | Notes |
|---|---|---|---|---|---|---|---|---|---|
| Wine Australia Export Label Image Search System (ELISS) | Wine Australia (Australian Grape and Wine Authority) | Every wine label approved for export since 2021-07-01: brand, product name, exporter/licensee, approval date, front/back label images, and declared vintage/variety/GI where shown. | Public directory; bulk-access request pending with Wine Australia | weekly | not yet | — | high | off | Off — Wine Australia declined bulk access and does not authorise automated collection (2026-09-01); no ingest |
off: not currently ingested — see its note for why. pending: never read, and still within one cadence of being enabled. fresh: read within the last cadence window. due: last read between one and two cadence windows ago. stale: never read and more than one cadence past being enabled, or last read more than two cadence windows ago.
Geo
| Source | Publisher | What we take | License | Cadence | Last read | Records | Confidence | Status | Notes |
|---|---|---|---|---|---|---|---|---|---|
| American Viticultural Area boundaries | UC Davis Library (AVA Digitizing Project) | Boundary polygons, CFR index, state/county and parent/child relations for every established AVA. | CC0-1.0 | monthly | 64,138 | high | fresh | — |
off: not currently ingested — see its note for why. pending: never read, and still within one cadence of being enabled. fresh: read within the last cadence window. due: last read between one and two cadence windows ago. stale: never read and more than one cadence past being enabled, or last read more than two cadence windows ago.
Reference
| Source | Publisher | What we take | License | Cadence | Last read | Records | Confidence | Status | Notes |
|---|---|---|---|---|---|---|---|---|---|
| Vitis International Variety Catalogue (VIVC) | Julius Kühn-Institut (JKI) | Prime grape variety names, VIVC numbers, synonyms, species, berry colour and country of origin; the reference for grape normalization. | Public database; cite VIVC (no bulk redistribution) | quarterly | — | high | fresh | — | |
| Wikidata (entity reconciliation) | Wikimedia Foundation | QIDs and identity links for grape varieties (via VIVC ID P3904), appellations (via eAmbrosia file number P9854) and producers (via official website domain P856); feeds identifier_mappings and JSON-LD sameAs. No fact overwrites. | CC0-1.0 (Wikidata public domain dedication) | monthly | — | medium | fresh | SPARQL endpoint; polite rate; exact identity joins only — fuzzy candidates are held for human review, never auto-linked |
off: not currently ingested — see its note for why. pending: never read, and still within one cadence of being enabled. fresh: read within the last cadence window. due: last read between one and two cadence windows ago. stale: never read and more than one cadence past being enabled, or last read more than two cadence windows ago.
Rules
| Source | Publisher | What we take | License | Cadence | Last read | Records | Confidence | Status | Notes |
|---|---|---|---|---|---|---|---|---|---|
| BLE Produktspezifikationen (German g.U./g.g.A. wine rulebooks) | Bundesanstalt für Landwirtschaft und Ernährung (BLE) | Produktspezifikationen for German protected wine names: permitted varieties, yields, alcohol, aging, zones, styles. | German official works (§5 UrhG) — confirm on site legal notice | quarterly | — | high | fresh | RECON 2026-09-29: VERIFIED (BLE index, >=46 PDFs); enabled with the working adapter (Task 8). | |
| BOE / Spanish DO appellation rulebooks (pliegos de condiciones) | Boletín Oficial del Estado (BOE) / Spanish DO regulatory councils | Pliegos de condiciones (specification documents) for Spanish DO/DOP wine appellations: permitted varieties, yields, min alcohol, aging requirements, zone boundaries. Not ingested. | Public documents; no national bulk source identified — deferred | quarterly | not yet | — | high | off | Off — deferred — eAmbrosia singleDocument is null for legacy wine GIs (recon 2026-09-01); no Spanish national bulk source identified; ES rules await a national route or eAmbrosia backfill |
| eAmbrosia product specifications (EU wine GI rulebooks) | European Commission — eAmbrosia register (documents hosted by member-state authorities) | Product specification (single document / cahier / disciplinare) PDF for every registered EU wine PDO/PGI that publishes one via eAmbrosia — permitted varieties (with colour scope and other-variety allowances), yields, alcohol minima, aging, zones, styles. | Official documents; per-host reuse terms (see adapter header) | quarterly | — | high | fresh | spec 2026-09-29 appellation conformance §5.2; skips GIs already covered by masaf/inao | |
| INAO cahiers des charges (French appellation rulebooks) | Institut National de l'Origine et de la Qualité (INAO) | Cahiers des charges (specification documents) for French wine AOC/IGP appellations: permitted varieties, min/max %, max yield (hl/ha), min natural alcohol, min aging (months, vessel), zone communes, styles allowed — one PDF per appellation, discovered via data.gouv.fr CSV + fiche-produit crawl. | Licence Ouverte 2.0 (data.gouv.fr seed CSV); public documents (cahiers des charges) | quarterly | — | high | fresh | — | |
| MAPA pliegos de condiciones (Spanish DOP/IGP wine rulebooks) | Ministerio de Agricultura, Pesca y Alimentación (MAPA) | Pliegos de condiciones for Spanish DOP/IGP wines: permitted varieties, yields, alcohol, aging, zones, styles. | Spanish public-sector information reuse (Ley 37/2007) — confirm on site legal notice | quarterly | not yet | — | high | off | Off — RECON 2026-09-29: NOT FOUND — MAPA has no national pliego index (JS-rendered pages; pliegos fragmented across 17 Comunidades Autónomas; see docs/superpowers/recon/2026-09-29-es-de-rulebooks.md). Stays disabled; Spain covered by eambrosia_specs (20/149 GIs) until per-region adapters exist. |
| MASAF disciplinari (Italian appellation rulebooks) | Ministero dell'Agricoltura, della Sovranità Alimentare e delle Foreste (MASAF) | Disciplinari di produzione (production rulebooks) for Italian DOC/DOCG/IGT appellations: permitted varieties, min/max %, max yield (hl/ha), min natural alcohol, min aging (months, vessel), zone communes, styles allowed — bulk .7z archives of per-appellation PDFs. | Italian PA open reuse (confirm IODL/CC-BY on legal notice — see adapter header) | quarterly | — | high | fresh | — | |
| US appellation labeling rules (27 CFR 4.25, Part 9; CA/OR/WA statutes) | TTB / eCFR; California, Oregon, Washington legislatures | What an AVA, county or state appellation claim legally implies: minimum percentage from the named area and conjunctive-labeling requirements. | Public domain (US federal and state law) | yearly | — | high | fresh | — |
off: not currently ingested — see its note for why. pending: never read, and still within one cadence of being enabled. fresh: read within the last cadence window. due: last read between one and two cadence windows ago. stale: never read and more than one cadence past being enabled, or last read more than two cadence windows ago.