Pull directional Google Trends data programmatically while accounting for the lack of a stable public API

Steps

1. Check whether Google's official Trends API (launched in alpha in mid-2025) covers your use case and request access, since it is the only Google-sanctioned programmatic path but remains limited-access with restricted endpoints and quota
2. If alpha access isn't available, treat unofficial libraries (e.g. pytrends or community forks) as reverse-engineered wrappers around the public trends.google.com website, not a supported API
3. Note that the original pytrends project was archived by its maintainers in 2025 and is no longer patched against upstream changes, so evaluate actively-maintained forks before depending on it
4. Rate-limit and randomize request timing aggressively when using unofficial methods, since Google can rate-limit or block IPs making automated Trends requests with no formal quota to negotiate
5. Treat all Trends values as relative/normalized search-interest indices (0-100 scaled per query), never as absolute search volume, regardless of which method sourced the data
6. Cache pulled data and re-fetch only periodically, both to reduce block risk and because Trends data for recent, low-volume periods can be revised after the fact