OpenDataSoft Scraper

Extract datasets and records from any of 3,000+ OpenDataSoft public portals — public.opendatasoft.com, data.paris.fr, data.economie.gouv.fr, BODACC, BOAMP, and thousands more. ODSQL where/select/group_by/order_by, facets, auto-pagination. No API key, no browser.

What does OpenDataSoft Scraper do?

OpenDataSoft (now Huwise) powers the open data portals of 3,000+ organizations — from the City of Paris and the French Ministry of Economy to the BODACC legal notices registry, RTE, and thousands of local governments worldwide. Every portal runs the same Explore API v2.1.

This actor is the generic reader:

• Catalog search — search a portal's dataset catalog by keyword, list all datasets, get metadata (title, description, theme, publisher, license, record count).
• Dataset metadata — full schema for any dataset: field names, types, labels, descriptions.
• Records — pull actual data rows with ODSQL where, select, group_by, order_by, and refine (facet filters). Auto-paginates up to 10,000 rows per dataset.
• Facets — list facet definitions and their distinct values.

No API key required. Anonymous access on public.opendatasoft.com allows up to 10 million calls per day. Other portals have their own quotas.

Output fields

Dataset catalog / metadata

Field	Description
datasetId	Dataset identifier
title / description	Dataset title and description
theme / keyword	Classification tags
publisher / license	Publisher and license
recordsCount	Number of rows
modified	Last modification date
fields	Schema (metadata mode): { name, label, type, description }
url	Direct portal link

Record

Every row is emitted as-is from the dataset, with three internal fields added:

Field	Description
recordType	record
portal	Portal hostname
datasetId	Source dataset
…all dataset columns	Whatever columns the dataset defines

Facets

Field	Description
facets	Array of facet names available on the dataset

Who is it for?

• Civic tech builders pulling permits, transit, environmental, and municipal data at scale
• Data journalists aggregating datasets across multiple French and international portals
• Market researchers combining business registries, legal notices, and tender data
• Compliance teams monitoring BODACC insolvency and BOAMP public procurement feeds
• Real estate analysts sourcing building permits, zoning, and property data
• Data scientists building pipelines across thousands of open data sources

Pricing

$0.30 per 1,000 results. No subscription.

Results	Cost
100	$0.03
1,000	$0.30
10,000	$3.00

How to use it

1. Pick a Mode.
2. Set the Portal hostname (e.g. public.opendatasoft.com).
3. For records: enter Dataset ID and optionally where, select, group_by, order_by, refine.
4. Set Max Items (default 10000).
5. Click Start.

Output example

{
  "recordType": "record",
  "portal": "data.paris.fr",
  "datasetId": "les-arbres",
  "id": "12345",
  "arrondissement": "PARIS 5E ARRDT",
  "genre": "Alignement",
  "espece": "Platanus x hispanica",
  "hauteur": 15.5,
  "circonference": 210,
  "annee_plantation": 1950,
  "adresse": "BOULEVARD SAINT-GERMAIN",
  "geo_point_2d": { "lat": 48.8501, "lon": 2.3448 },
  "scrapedAt": "2026-09-26T12:00:00.000Z"
}

Technical details

• Source: OpenDataSoft Explore API v2.1 — https://{portal}/api/explore/v2.1.
• No API key required for public datasets. Anonymous access is allowed on all portals.
• Rate limits — 10M calls/day on public.opendatasoft.com. Other portals have their own quotas communicated via X-RateLimit-* headers. The actor backs off on 429.
• ODSQL query language — same syntax across all endpoints. Supports where, select, group_by, order_by, refine, and the search() full-text function.
• Pagination — limit per request caps at 100. Offset pagination caps at 10,000 records per dataset.
• 3,000+ portals — any OpenDataSoft instance works: public.opendatasoft.com, data.paris.fr, data.economie.gouv.fr, bodacc-datadila.opendatasoft.com, boamp-datadila.opendatasoft.com, opendata.paris.fr, and thousands more.
• No browser, no proxy — pure REST JSON.

Known limits

• Records per request cap at 100. Offset pagination caps at 10,000 rows per dataset. For deeper extraction, use the dataset's export endpoints or filter by date to split queries.
• Anonymous quotas vary by portal. public.opendatasoft.com allows 10M calls/day. Smaller portals may have lower limits. The actor respects X-RateLimit-* headers.
• Some datasets are restricted. Public datasets work anonymously; authenticated-only datasets return 403.
• Field names are portal- and dataset-specific. Each dataset has its own schema. The datasetId and portal fields are attached to every record for downstream filtering.
• No PDF or file attachments. The actor returns structured record data. Dataset file exports (CSV, GeoJSON, Shapefile) are separate endpoints.

FAQ

Do I need an API key? No. Anonymous access works for all public datasets.

Do I need a proxy? No. Datacenter IPs work.

How do I find a dataset ID? Browse the portal, open a dataset, click API — the dataset_id is in the endpoint URL.

What's the difference between catalog and records mode? Catalog returns dataset metadata (title, description, record count). Records returns the actual rows.

What ODSQL functions are available? search(), count(), sum(), avg(), min(), max(), year(), month(), day(), distance(), and standard comparison operators.

How do I export data? After a run, go to Storage → Export as JSON, CSV, Excel.

Support

Open an issue on the Actor's page for bugs or feature requests.