1d ago · 6 min read · When we shipped the first version of AI-generated replies for HelperX, each reply cost us about \(0.011 in API spend. That sounds tiny until you multiply by 30 replies per slot per day times 200 activ
Join discussion
2d ago · 30 min read · A week ago, I saw this tweet: I had just shipped SupportMesh, a multi-tenant AI support platform built on Mastra, so I had opinions from production. I liked the .dowhile() loop, the typed step schem
Join discussion
4d ago · 6 min read · Originally published in Japanese on Zenn. On Hashnode, set this as the canonical in the publish panel: open Article Settings, find "Are you republishing?", and paste the Zenn URL under "Add Original U
Join discussion