Skip to content

Developers

Build on the catalog

A public, read-only REST API over every dataset in the library — the same documentation the site shows, as JSON.

Quickstart

Your first request in one line

Search, filter by license, and fetch full documentation for any dataset. Responses are JSON; no key is needed.

Base URL
https://datazimuts.com
Authentication
None — the catalog API is public and read-only.
curl
curl "https://datazimuts.com/v1/datasets?q=unemployment&commercial_use=yes&per_page=5"
JavaScript
const res = await fetch("https://datazimuts.com/v1/datasets/facets");
const { sources, tags, commercial_use } = await res.json();
console.log(sources.length, "sources,", commercial_use.yes, "datasets cleared for commercial use");
Python
import requests

BASE = "https://datazimuts.com"
hits = requests.get(f"{BASE}/v1/datasets", params={"q": "inflation", "per_page": 5}).json()
for d in hits["datasets"]:
    print(d["id"], d["rows"], d["commercial_use"])

doc = requests.get(f"{BASE}/v1/datasets/{hits['datasets'][0]['id']}").json()
print([c["name"] for c in doc["schema"]])

Licensing

Respect each dataset’s license

Every dataset carries its publisher’s license and a commercial-use clearance. Filter with commercial_use and always check the license before reuse.

  • yesCommercial use is permitted by the license (attribution may still be required).
  • unclearThe license terms are unclear — review them before commercial use.
  • noThe license prohibits commercial use. Documentation stays open; downloads return HTTP 403.

Reference

Endpoints

Generated from the live OpenAPI specification, so it always matches the running API. Endpoint descriptions come from the specification in English. Try requests interactively in the interactive API console

  • GET/v1/datasetsList Datasets

    Search and filter the catalog, with pagination (page is 1-based).

    q matches titles, descriptions, tags, themes and source names. commercial_use is yes, no or unclear. sort is quality, updated, rows or title; omitted, results keep a stable default order. per_page is capped at 100.

    q
    query · string
    source
    query · string
    tag
    query · string
    commercial_use
    query · string
    sort
    query · string
    page
    query · integer
    per_page
    query · integer
  • GET/v1/datasets/{id}Get Dataset

    Full documentation for one dataset: schema, sample rows, license, attribution and Croissant (MLCommons) JSON-LD metadata.

    id*
    path · string
  • GET/v1/datasets/statsCatalog Stats

    Catalog totals: datasets, rows, sources and the latest refresh time.

  • GET/v1/datasets/facetsDataset Facets

    Every filter value in the catalog, with counts.

    Sources, tags, themes and commercial-use clearance across the whole catalog — use these values with the source, tag and commercial_use filters of the listing.

  • GET/v1/datasets/{id}/downloadDownload Dataset

    Download a dataset's sample rows (up to 20) as a CSV or JSON attachment.

    Returns 403 when the dataset's license prohibits commercial use (commercial_use = no); its metadata, schema and sample rows stay readable through the detail endpoint.

    id*
    path · string
    format
    query · string

Agents & LLMs

Machine-readable by design

  • /llms.txt

    A plain-text index of every dataset with links, following the llms.txt convention.

  • JSON-LD · Croissant

    Each dataset page embeds schema.org Dataset and MLCommons Croissant metadata for search engines and ML tools.

  • /openapi.json

    The full OpenAPI 3 specification, for client generators and tool-using agents.