Hey! I built call-latest to kill async race conditions & heavy boilerplate. It’s the canonical async-safety primitive for modern AI apps.
Centering around createSmartSearch, it’s a production-grade orchestration engine in a tiny ~1.3KB core:
• Adaptive Debounce
• Backtrack LRU Cache + SWR
• Edge-Native Distributed Cancellation
• Delta Merging for bandwidth saving
• AI-friendly (Cursor/Copilot/Claude ready)
Zero-dependency, TS-first. Check the docs & benchmarks on GitHub!
