Skip to content

Lexicon · reference data

French agricultural data, linked together

Land registry, CAP plots, plant protection products, weather, businesses, agronomic knowledge: Lexicon brings together 623M public records and links them with shared keys. A commons of the OSFarm association, open to all.

linked records
623M
datasets
54
data providers
33
datasets open to all
46

Who does what

Lexicon is a commons of the OSFarm association, and Ekylibre SAS is its main contributor. We produce 16 of its datasets, such as crop management plans, intervention templates and the agricultural chart of accounts.

7 families of data

INSEE, IGN, ANSES, Météo-France, Agence Bio, ASP and many more: each dataset is taken from its source, kept up to date and linked to the others.

  • Land and territory

    11 datasets · 289M rows

    Cadastral parcels · Hydrography · Cadastral owners · CAP declared plots (LPIS) · Land prices …

  • Climate

    1 dataset · 221M rows

    Weather observations

  • Inputs and regulation

    8 datasets · 61M rows

    Plant protection product purchases · Authorised plant protection products · Active substances · Plant varieties · Fertilisers …

  • Businesses and subsidies

    11 datasets · 51M rows

    Businesses (SIREN) · CAP beneficiaries · Business links · Organic farming · Quality and origin labels …

  • Markets and economics

    5 datasets · 192K rows

    EU market prices · Reference prices · Agricultural chart of accounts · Fuel prices · Budget templates

  • Agronomic knowledge

    13 datasets · 399K rows

    Agricultural R&D · Agricultural productions · Open nomenclature · Product variants · Crop management plans …

  • Food systems

    5 datasets · 242K rows

    School and public canteens · Food safety inspections · Territorial food projects · Agricultural education · AgroEDI exchange standard

All datasets, with their source and licence

Four ways to use it

  • In the browser

    Every dataset can be browsed and filtered on lexicon.osfarm.org. The “Identify a plot” tool shows everything Lexicon knows about a point on the map.

  • Through the API

    Every page also exists in JSON and CSV: just add .json or .csv to its address.

  • In your database

    Each dataset is a versioned package you can load into your own PostgreSQL database.

  • Through an AI assistant

    Lexicon’s MCP server lets an assistant search and cross-reference the datasets, with their sources.

bash
curl https://lexicon.osfarm.org/phytosanitary/products.json
Response (excerpt)
{
  "title": "Produits",
  "table": {
    "columns": { "name": "Nom", "firm": "Firme", "type": "Type",
                 "active-compounds": "Principes actifs", "state": "État d'usage" },
    "rows": [
      { "name": { "@type": "Text", "value": "…" },
        "type": { "@type": "Text", "value": "PPP" },
        "active-compounds": { "@type": "List", "values": ["…"] },
        "state": { "@type": "Text", "value": "Autorisé" } },
      …
    ]
  },
  "credit": { "provider": { "value": "ANSES" }, "license": { "value": "Open Licence" }, … },
  "page": 1, "items-per-page": 150, "items-total": …, "total-pages": …,
  "formats": { … }
}

A paginated response, with the dataset’s source, date and licence. No key, no sign-up.API and MCP details

Open, sourced, traceable

Lexicon’s source code is open on GitHub. Each dataset keeps the licence of its source, shown in the catalogue with its date and version.

46 datasets are open to all. 8 are reserved for members of the OSFarm association, who access them with an API key.

Who it is for

  • Software vendors and agtech start-ups

    up-to-date reference data without having to collect it yourself.

  • Co-ops and farm organisations

    public data ready to cross-reference with your members’ data.

  • Researchers, teachers, local authorities

    a single base for studying an area or a supply chain.

  • Farmers

    indirectly, through Ekylibre and Aiffel, which use it to check a rate or identify a plot.

Need Lexicon inside your tools?

We help you connect the API, load the PostgreSQL packages or propose a new dataset.