A free, no-auth REST API for real-time AI news, model rankings, and trending tools.
No API key. No signup. Just hit the endpoints. Powered by whatstrending.ai.
curl "https://whatstrending.ai/api/articles?limit=3"- AI news feed — fresh, summarized, categorized AI/tech news, each linking back to the original source.
- AI model leaderboard — live rankings ranked by real OpenRouter token volume, with live pricing and context windows.
- Trending AI tools — a curated, auto-updated list of notable AI tools.
- Most-read — what people are actually reading right now.
All read-only GET endpoints. No authentication, CORS enabled, JSON responses.
Beyond JSON, key pages ship as clean text/markdown (CORS enabled, no key) — ideal for AI agents:
/models.md— the usage leaderboard as a markdown table./news.md— recent headlines with sources and links./compare/{slug}.md— any comparison page as markdown.
Also on the site, built from the same tracked data:
- Daily AI Pulse — daily briefing (top stories, model movers, rising repos, funding), archived per day.
- Rising repos — week-over-week star velocity from stored daily snapshots.
- Model Release Radar — new models as they appear on OpenRouter, with launch pricing and early adoption.
- AI Market Share — provider share of real token volume over time.
- Trending AI Repo badge — if your repo is on the trending index, embed it:
[](https://whatstrending.ai/repos/OWNER/NAME)
https://whatstrending.ai
None. Every endpoint below is public — no key, no header, no signup.
| Param | Type | Default | Description |
|---|---|---|---|
category |
string | — | Filter by category (see below) |
limit |
number | 50 | Max items to return |
offset |
number | 0 | Skip N items (pagination) |
curl "https://whatstrending.ai/api/articles?category=Tools&limit=2"{
"success": true,
"data": [
{
"title": "The Atlantic created a searchable database of the music used to train AI",
"originalTitle": "The Atlantic created a searchable database of the music used to train AI",
"slug": "the-atlantic-created-a-searchable-database-of-the-music-used-to-train-ai",
"link": "https://www.theverge.com/ai-artificial-intelligence/953183/the-atlantic-searchable-database-music-ai-training-data",
"source": "The Verge",
"category": "Tools",
"date": "2026-06-20T14:46:48-04:00",
"sourceType": "full",
"summary": "Atlantic reporter Alex Reisner uncovered four datasets of music being used to train AI models and made them fully searchable for the public...",
"coverage": 1,
"trendScore": 1,
"sources": ["The Verge"]
}
]
}Each item also carries coverage and trendScore (numbers indicating how widely the story is being reported and how fast it's trending) and sources (array of publication names covering the story). The date field is passed through from the source feed, so it may be ISO 8601 (2026-06-20T14:46:48-04:00) or RFC 2822 (Sat, 20 Jun 2026 16:39:57 +0000).
Categories are dynamic and reflect current coverage — commonly Industry, Tools, Research, Startups, Regulation, Models, Open Source.
curl "https://whatstrending.ai/api/articles/openai-is-bringing-on-some-big-guns-in-the-lead-up-to-its-ipo"Returns one article object in { "success": true, "data": { ... } }. Returns 404 with { "success": false, "error": "Article not found" } if the slug doesn't exist.
| Param | Type | Default | Description |
|---|---|---|---|
limit |
number | — | Max models to return |
curl "https://whatstrending.ai/api/models?limit=1"{
"success": true,
"data": [
{
"rank": 1,
"name": "DeepSeek V4 Flash",
"provider": "DeepSeek",
"score": 4676,
"scoreDisplay": "4.68T",
"tokens": 4676449991973,
"requests": 362209743,
"pricing": "$0.09/$0.18",
"context": "1M",
"category": "Open Source",
"change": "0",
"or": "deepseek/deepseek-v4-flash"
}
]
}The leaderboard is ranked by real OpenRouter usage. tokens is the trailing-window token volume routed to the model, requests is the request count, and scoreDisplay is the human-readable form of score (e.g. "4.68T"). pricing is $in/$out per million tokens and or is the OpenRouter model id.
Look up one model by id. The {id} is the name slugified — lowercased with spaces replaced by hyphens (e.g. DeepSeek V4 Flash → deepseek-v4-flash). It is not the or/OpenRouter id; passing that (which contains a /) will not match.
curl "https://whatstrending.ai/api/models/deepseek-v4-flash"{
"success": true,
"data": {
"rank": 1,
"name": "DeepSeek V4 Flash",
"provider": "DeepSeek",
"score": 4676,
"scoreDisplay": "4.68T",
"tokens": 4676449991973,
"requests": 362209743,
"pricing": "$0.09/$0.18",
"context": "1M",
"category": "Open Source",
"change": "0",
"or": "deepseek/deepseek-v4-flash"
}
}Returns 404 with { "success": false, "error": "Model not found" } if no model matches the slug.
curl "https://whatstrending.ai/api/most-read"{
"success": true,
"data": [
{ "slug": "laid-off-oracle-workers-tried-to-negotiate-better-severance-oracle-said-no", "views": 10 },
{ "slug": "my-yard-is-dying-so-i-made-an-app-for-that", "views": 10 }
]
}Each entry is { slug, views }, ordered by views. Use the slug with GET /api/articles/{slug} to fetch the full article.
curl "https://whatstrending.ai/api/awesome-data"{
"generatedAt": "2026-06-21T01:46:39.887Z",
"site": "https://whatstrending.ai",
"tools": [
{
"name": "Cursor",
"tagline": "The AI-first code editor",
"description": "AI-powered code editor built on VS Code with deep integration for code generation, editing, and chat.",
"url": "https://cursor.com",
"category": "coding",
"pricing": "freemium"
}
],
"comparisons": [
{ "a": "ChatGPT", "b": "Claude", "slug": "chatgpt-vs-claude" }
],
"models": [
{
"rank": 1,
"name": "DeepSeek V4 Flash",
"provider": "DeepSeek",
"score": 4676,
"scoreDisplay": "4.68T",
"tokens": 4676449991973,
"requests": 362209743,
"pricing": "$0.09/$0.18",
"context": "1M",
"category": "Open Source",
"change": "0",
"or": "deepseek/deepseek-v4-flash"
}
]
}This endpoint returns three collections:
tools— curated AI tools, each withname,tagline,description,url,category, andpricing(e.g.free,freemium,paid).comparisons— head-to-head pairs, each{ a, b, slug }. Theslugmaps to a comparison page (e.g./compare/chatgpt-vs-claude).models— the same model objects asGET /api/models(same 12-field schema).
Note:
/api/awesome-datareturns{ generatedAt, site, tools, comparisons, models }directly (not the{ success, data }envelope used by the other endpoints).
Most endpoints use a consistent envelope:
{ "success": true, "data": <payload> }
{ "success": false, "error": "Description" }All responses send Access-Control-Allow-Origin: *, so you can call the API directly from the browser.
Beyond the JSON API, the site exposes static, machine-readable surfaces meant for LLMs, agents, and feed readers:
| Surface | Format | What it is |
|---|---|---|
/llms.txt |
text | LLM-friendly site summary and intent → page map |
/news.md |
Markdown | Latest AI news, summarized, with source links |
/models.md |
Markdown | Model leaderboard as a Markdown table |
/feed.xml |
RSS 2.0 | News feed for RSS readers |
/sitemap.xml |
XML | Full sitemap |
These are great if you want a quick, parse-free snapshot without hitting the JSON endpoints.
There's no hard rate limit today, but please be reasonable — cache responses where you can and don't hammer it. The API is offered free; abuse spoils it for everyone.
News items are summarized by AI from publicly available sources; every item includes a link and source pointing to the original publication — follow those for the full story and credit. Model rankings and tool listings are aggregated and updated automatically. Free to use; a link back to whatstrending.ai is appreciated.
Runnable snippets live in examples/ — curl, JavaScript (fetch), and Python (requests).
MIT — do whatever you want, no warranty.
Built by whatstrending.ai · the pulse of AI, summarized.