1.0 (2026-08-21) — Large runs now finish and return their results
Runs covering many products or reviews could end without returning the data they had already gathered
Results are now written in batches, so a large run completes and returns everything it collected
Progress is now shown while results are being written, instead of appearing stalled
1.0 (2026-06-04) — Faster multi-product runs with live progress
Product and ASIN batches now fetch in parallel instead of one at a time, so large runs finish much faster
The status now shows live fetch progress instead of appearing to stall on big batches
1.0 (2026-05-23) — Fix: rare hang during cross-run watchlist memory access
1.0 (2026-05-23) — Clearer failure messages when Amazon fetch fails
1.0 (2026-05-23) — Fixed missing run summary, manifest, and incident-queue artifacts
1.0 (2026-05-23) — Bug fix: prevent rare run timeouts during finalisation
1.0
Initial release. Persistent Amazon catalog operations intelligence: a ranked incident queue across products, reviews, and sellers with price + BSR trajectory, buybox/seller change detection, deterministic defect emergence from review themes, comp-set position, seller intelligence, market regime, and persistent monitor-mode memory across runs. Six modes (search, products, reviews, compset, monitor, seller), cross-marketplace support with country-pinned residential proxy, and drop-in compatibility with the junglee/Amazon-crawler input shape and field set.
