The Crosswire open dataset
Every outlet Crosswire tracks — 907 newspapers, broadcasters, agencies and digital outlets across 186 countries — published as a single JSON file under CC BY 4.0. No signup, no API key, no rate limit: it is a static file served from the same edge as the site, and it is regenerated with every release.
What's in it
The file opens with the licence, the attribution string, the generation date and the scale used for editorial leaning, then lists all 186 countries. Each country carries its Crosswire URL and its outlets; each outlet carries six fields.
- id
- Stable slug, also the outlet's Crosswire URL: /outlet/{id}
- name
- The outlet's own name, in its usual Latin-script form
- homepage
- Canonical homepage URL
- lean
- left | center | right — approximate editorial orientation
- type
- newspaper, tabloid, business daily, magazine, public broadcaster, tv news, news agency, digital outlet or news portal
- description
- A short English description of the outlet
One outlet, in full
{
"id": "lemonde",
"name": "Le Monde",
"homepage": "https://www.lemonde.fr",
"lean": "center",
"type": "newspaper",
"description": "Le Monde is one of France's most influential daily newspaper…"
}Real output for Le Monde, trimmed only in the description field.
How to cite it
CC BY 4.0 asks for credit, a link and an indication of any changes. In practice that means a line like this, wherever the data appears:
News outlet directory from Crosswire (crosswire.online), CC BY 4.0. Retrieved 2026-07-28.
If you publish an analysis based on the leaning labels, please say that they are approximate orientations rather than measurements, and link to the methodology so readers can see how they were arrived at. Corrections are welcome at [email protected].
What it is not
This is a directory, not a bias measurement. The lean field is a coarse three-way grouping meant to make cross-perspective reading easy — it is drawn from widely cited media-bias assessments and each outlet's reputation, and it is within-country: a title we mark Center in one country may sit well left or right of a Center title in another, because the spectrum itself differs. It carries no claim about accuracy, quality or trustworthiness, and where no confident rating was available an outlet is shown as Center.
The dataset also contains no article text and no headlines. Crosswire renders each outlet's live front page through a proxy but never stores it, so there is no historical record here to query — only the directory itself.
Other machine-readable formats
/llms.txt is a short index of the site for AI assistants; /llms-full.txt is the whole corpus as plain text, grouped by country and ordered left to right. Licensing terms for automated use are published as RSL 1.0. Release history is available as an Atom feed.