Research
Articles are built from pages we actually read, and you can check every one.
How it works
We search for relevant sources, fetch a bounded set of pages, and extract the passages that support each claim. Every source is stored with its URL, title, retrieval time and the excerpt used.
What we record
Source URL, page title, when we read it, the supporting excerpt, the publication date when the page exposes one, and the retrieval status. A source we could not read is kept with the reason — not silently dropped.
What we will not do
We do not reproduce competitor articles, manufacture first-hand experience, or invent statistics. If a claim has no supporting source, the evidence check removes it and the article goes to your review.
Respecting other sites
We honour robots.txt, limit how many pages we read per domain, and identify ourselves honestly. A site that asks us not to read a page is recorded as disallowed and contributes nothing.