Streaming generation and cloud storage: feeds up to 100 MB
Feedflock now renders feeds product by product straight to cloud storage instead of building them in memory, so large catalogs generate as comfortably as small ones.
Until now, Feedflock built each feed in memory and wrote the file out at the end. That works for a few thousand products, but a large catalog (variants, metafields and gallery images included) could outgrow what a single run can hold, and you only found out when the generation failed.
Generation now streams. The renderer walks your catalog product by product, writes each item straight to cloud storage (S3/R2) and never holds more than a slice of the feed at once. A 100 MB feed file generates as comfortably as a small one, and the bulk export behind it spools to disk instead of memory.
Password-protected feeds get private storage with presigned URLs on top, so a Basic-Auth feed is never publicly reachable, not even by guessing the file address.