The New York Times API, RSS and sitemap
Every article The New York Times publishes, as JSON, through an open API with no key and no signup. 17,116 of their articles are indexed right now, refreshed hourly.
United StatesEnglishSpanish#385 by volume
Four ways to pull nytimes.com
All four are ours, all four are live, and none of them needs a key.
JSON Search this publisher
Every article we hold from nytimes.com, newest first. Add q= to search inside.
curl -s "https://freenewsapi.ai/v1/search?host=www.nytimes.com&size=20"open it →RSS Subscribe in any reader
The same articles as an RSS 2.0 feed. Works in Feedly, Inoreader, Thunderbird — anywhere.
curl -s "https://freenewsapi.ai/v1/rss?host=www.nytimes.com&size=50"open it →JSON Feed For apps
JSON Feed 1.1, the same content in a format apps parse natively.
curl -s "https://freenewsapi.ai/v1/feed.json?host=www.nytimes.com&size=50"open it →Trimmed For agents
Only the fields an agent needs — title, url, date, language — so the response stays small.
curl -s "https://freenewsapi.ai/v1/search?host=www.nytimes.com&size=20&fields=title,url,published_at,lang"open it →Call the API
A live request. Paste it into a terminal and The New York Times headlines come back as JSON.
curl "https://freenewsapi.ai/v1/search?host=www.nytimes.com,open.nytimes.com&size=5"This page covers 2 hosts of nytimes.com as one publisher: www.nytimes.com, open.nytimes.com. The API accepts them comma-separated in host.
Publishing volume
Articles per day over the last 65 days. The busiest hour is 09:00 UTC — that is when the newsroom pushes its main batch.
The 6 hatched days are gaps in the upstream news archive, not quiet days at The New York Times: the whole corpus dips on those dates. They are excluded from the median above.
curl "https://freenewsapi.ai/v1/stats?host=www.nytimes.com,open.nytimes.com&top=10"Publishing rhythm
When The New York Times actually publishes, by hour of day (UTC) and by day of week — measured from 17,116 timestamps, not from a schedule they advertise.
Sections
Sections come from the publisher's own markup, present on 1% of their articles.
Languages
The site identifies itself under The New York Times, NYT > Science, Natural 20, Daring Fireball, NYT Open - Medium — separate editions publishing under one domain.
| Code | Language | Articles | Share |
|---|---|---|---|
en | English | 17,073 | 99.7% |
es | Spanish | 33 | 0.2% |
Latest articles
The 18 most recent The New York Times articles in the index — the same rows the API returns, rendered for humans.
Harvest it all
Page through the archive by date. from and to accept
YYYY-MM-DD or relative values such as now-7d.
curl "https://freenewsapi.ai/v1/search?host=www.nytimes.com,open.nytimes.com&from=now-7d&size=100"To poll for new material, keep the timestamp of the newest article you saw and ask only for what came after it:
curl "https://freenewsapi.ai/v1/search?host=www.nytimes.com,open.nytimes.com&from=now-1h&sort=date_asc"Article bodies come with full_text=true. No key, no quota tier, no
header — see the search endpoint for every parameter.
Metadata quality
How well The New York Times marks up its own pages — useful if you are deciding whether to parse them yourself.
Crawl lag — the delay between publication and the article reaching this index — is 9 min at the median, 1058 minutes at the 90th percentile.
Similar publishers
Comparable outlets in the same language. Those without a page here yet link straight to their slice of the API.
FAQ
Does The New York Times have an official API?
Not a public one. This endpoint is the practical alternative: their articles as structured JSON, with title, lead, publication time, section, byline and a link back to the original.
How do I get their RSS feed?
Most publishers expose one at /rss or /feed, but feeds carry
only the last few dozen items and drop them within hours. This index keeps 17,116 of
their articles and lets you filter by date, language and section.
How many articles does The New York Times publish per day?
330 on a median day, measured across 65 days.
Can I use this commercially?
The API is free and unauthenticated. What you may do with the content is between you and the publisher: we return metadata, a lead paragraph and a link, not a licence.
