What PubTrace does, and how

Here's an illustrative example of a common pattern: a package launches, gets a burst of attention, then settles into a quieter, steady rhythm. pub.dev's UI only shows a rolling 30-day count. Once that launch spike ages out of the window, the number looks like it "dropped". It didn't; the historical downloads still happened. Same package, two ways:

Illustrative example, not a real package's data.

PubTrace fetches pub.dev's own public API, live, and sums the 52 weeks of download data it already publishes but doesn't surface in its UI. It shows you the whole year pub.dev is already tracking, not a new metric it invented.

1. You search a package, say dio, the ecosystem's most recognizable HTTP client.
2. PubTrace calls pub.dev's public API live; no caching layer hides where the numbers came from beyond a 1-hour refresh.
3. It sums the 52 weekly values and charts them. This is that chart, right now:

• No database of record. Nothing is stored between visits.
• No accounts, no signup, no tracking.
• No numbers older than 12 months. pub.dev's public API only exposes the trailing 52 weeks; anything before that genuinely isn't retrievable, by anyone, from this API.