# Changelog of Product Hunt Launch Pricing & Website Intel (`datagrit/product-hunt-launch-pricing-intel`) Actor

- **URL**: https://apify.com/datagrit/product-hunt-launch-pricing-intel/changelog.md
- **Full Actor documentation**: https://apify.com/datagrit/product-hunt-launch-pricing-intel.md

## Changelog

### 0.3

- A leaderboard page that Product Hunt occasionally answers with HTTP 200 but without its embedded data is now fetched once more before the run counts it as unreadable, so a single bad response no longer fails the whole run.
- The closing status message now counts one billed record per launch instead of two.

### 0.2

- Product Hunt answers HTTP 403 to every datacenter address, which failed the platform verification run. Product Hunt requests (leaderboards, product pages) now go through the Apify residential proxy by default (one retry on a new address after a 403); product websites are fetched directly. Platform run: 17 launches in 120 s, pricing parsed for 10 of 17 websites, total cost 0.0039 USD (residential 0.00014 GB).

### 0.1

- Initial release: daily, weekly, monthly and yearly Product Hunt leaderboards, single product lookup, resolved product website, pricing page extraction (plans, lowest paid price, pricing model), topic, vote and pricing model filters, only-new mode.
- Reliability: the run now fails instead of returning rows with empty pricing when no product website returns a readable page, or when pricing cannot be read on any of at least 3 reachable websites. It also fails when every leaderboard page and product page requested together is unreadable. The status message always reports "Pricing parsed for X of Y websites" with the reasons for the rest, and the closing sentence names only the filters that dropped launches.
- Manual input checks: typical (daily leaderboard, 12 launches, pricing parsed on most sites), edge (one product page by slug, weekly leaderboard with a topic filter and pricing off), no results (minimum votes above any launch gives one status row, not billed).