Kiro started as an experiment: "What if a task manager actually understood project structure instead of just storing todos?" The hardest part wasn't AI—it was building the scheduling engine underneath it. I ended up writing: • a custom virtualized timeline capable of handling multi-year schedules • a DAG-based dependency engine with cycle detection • optimistic state management with transactional rollback • timezone-safe scheduling using UTC day normalization • AI agents that manipulate real project state instead of simply chatting My goal was to make planning feel instantaneous, even when the underlying scheduling logic is fairly complex. I'd love feedback on the UX, architecture, or anything that could make it better.

Kiro

Plan complex projects with AI and zero-lag scheduling

Plan complex projects with AI and zero-lag scheduling