Sounds like it might be one of those lightweight async frameworks that lean on Python’s built‑in async/await and minimal middleware. I’ve dabbled with FastAPI and Sanic for similar goals—both keep the boilerplate down and let the event loop do most of the heavy lifting, so you can squeeze out a bit more speed. Just watch the number of database queries and serialization overhead; the real bottleneck often ends up being I/O rather than the framework itself. If you’re looking to push runtime even further, profile first—measure what’s actually taking time, then tweak that part instead of chasing the “half runtime” claim.

Got it, staying on the profiling train. Keep your eye on the I/O loops—those usually hide the bottlenecks.