public API · Cloudflare AI · no key

The whole library, programmatically.

Every dataset on this site ships as static JSON from Cloudflare’s edge — no key, no login, no bill. The fact-check engine scans every source for or against a claim and runs in your browser, so the whole thing works offline too. An optional label header helps you keep your own calls organized.

GET /data.json · GET /news_archive.json · GET /v1/sources.json · fact-check demo below

1 · no key required

Open by design

These are plain static files on Cloudflare’s edge — there is nothing to authenticate. If you want to tag your own requests, set an optional label below and it rides along as an x-api-key header purely for bookkeeping. The demo runs fully in your browser against the same static files this site ships.

your request label (optional)
No key needed. Labeling is optional and stays in your browser.
2 · endpoints

What you can call

GET
/data.json
Full hardcoded ledger — gas, debt, tariffs, markets, labor, accountability, world. Every figure dated and source-tagged.
GET
/news_archive.json
Embedded article archive — 400+ articles with full archived text (body[]) aggregated at build time, deduped across two feeds.
GET
/news_archive.json?tag=trump
Filtered archive — supports cat, tag, feed (newsapi / currents), and q for keyword search across title, description, and body.
POST
/v1/fact-check
Body: {"claim":"...","scan":"archive"}. Keyword + number-matches the claim against the full dataset and every source, returning SUPPORT / PARTIAL / NO DATA verdicts with source tags. Mirrors the in-browser demo below.
GET
/v1/sources.json
The bias & leaning index used by the Tools page (AllSides-style scale) so clients can weigh a source alongside its data.
3 · live fact-check demo · Cloudflare AI, no key

Paste a claim. Get evidence.

claim

verdict

Results appear here. The engine scans the full archive and source list, then marks each match SUPPORT, CONTRADICT, or NO DATA.
4 · code samples

Call the API


        
Ready. Click "Run the demo" to pull data.json straight from Cloudflare’s edge.

      

      

        
Run it here

        
        
Output from the keyed request lands here.

      

    

  


  
This is the same data the Explore page renders, exposed as endpoints — no key, free forever. Articles were aggregated at build time from upstream feeds, stripped to plain text, and baked into news_archive.json; the site makes no live upstream API calls at page-load, so it keeps working even when the feeds rate-limit or vanish. The fact-check demo is the same engine the Tools page uses: it runs client-side with a Cloudflare-prompted local model, and falls back to the built-in keyword scanner when your browser blocks the request.