Methodology

How ChangeRadar finds, verifies, and timestamps every change.

Pricing & rate-limit changes

For each tracked tool we capture a normalized snapshot of its public pricing page on a fixed schedule and store it with a content hash. When the hash changes, we compute a line-level diff and record exactly what was added or removed — building a permanent, timestamped history that the vendors themselves don't publish.

Deprecations & breaking changes

We read the official changelog/RSS feeds of major developer platforms and surface the items that signal a deprecation, sunset, end-of-life, or breaking change. Each item links back to the authoritative source.

How we collect data — responsibly

• Only public, logged-out pages and official feeds — no accounts, no circumvention.
• We identify our crawler with a contactable User-Agent and rate-limit politely.
• We store facts (prices, limits, dates) and short excerpts, and link to the source.

Freshness

Sources are re-checked on a recurring schedule (every few hours). A change typically appears within hours of the vendor publishing it. Spotted an error or a source we should add? Get in touch from the homepage.

Analytics & privacy

Our site analytics are first-party and privacy-preserving by construction:

• No cookies and no consent banner — we set nothing on your device, so there is nothing to consent to.
• No IP address is ever stored. To count unique visitors without cookies we derive a one-way hash from a daily-rotating secret; the raw IP and user-agent exist only for that instant and are immediately discarded.
• No cross-day or cross-site tracking. The hashing secret is replaced (and the old one deleted) every night, so yesterday's visitors can't be re-identified today.
• Country-level geography only — we look up the visitor's country in a local database and keep just the 2-letter country code. No city, region, or coordinates.
• We honor Do Not Track and Global Privacy Control — if your browser sends either, we record nothing at all.
• Bot and datacenter traffic is excluded. Requests from hosting, cloud and VPN networks (looked up in a local IP-to-network database) are not counted as visitors — only an anonymous daily total of excluded requests is kept, so our traffic numbers describe humans.
• Every metric is computed and stored in our own database — no third-party analytics service, no tracking pixels.

IP-to-country and IP-to-network lookups use the DB-IP Lite databases, licensed under CC BY 4.0.