Public release log for FairCrawl. We ship continuously; the highlights live here. Subscribe to the changelog feed at faircompany.ai/changelog.
The scrape cascade gained an unlock step for heavily bot-protected pages — CAPTCHA walls, fingerprint checks, session-strict SPAs. It's billed only when it succeeds; a failed unlock still costs $0. Several long-blocked target classes now come back with real content.
The typed catalog nearly tripled — 63 to 185 active endpoints across 18 platforms. Every endpoint was verified against live traffic before activation, and the catalog counts only what you can actually call. Google Maps joined as a paid vertical: search, details, reviews.
Web-scale search that remembers what it already showed you. Schedule a query, get dedup across runs, LLM relevance judging, and a webhook the moment something new appears. Full SDK, CLI, and MCP support.
One call fans out from SERP to scraped pages. The agent endpoint hit GA. And extract-schema recipes now self-heal: element fingerprints relocate your fields when a site redesigns, so extractions survive layout changes.
YouTube and TikTok transcripts — captions first, tiered speech-to-text when captions don't exist. Deep pagination on X, Instagram, and TikTok history walks that used to stall on looping cursors.
Schedule a crawl, get a diff of what changed since last run, receive a completion webhook. The foundation the Monitor product built on.
Every Fair product portal now serves the standard page set: /docs, /api, /about, /terms, /privacy, /changelog. Cuts onboarding friction for new scrape users hitting these routes from a search engine.
Every scrape now passes through the shared Fair quality gate. Failed operations are voided — you're charged $0 when something doesn't deliver. Receipts show the gate decision in the cost-receipt JSON.
@fair/sdk ships a typed client for FairCrawl
with auto-retry, hold/settle handling, and exhaustive
type inference for every endpoint.
FairCrawl opened to the public. Cost-plus pricing live from day one. Full feature set from your first scrape.
Older entries archived. For the full timeline, see github.com/faircompany.