I'm the author of DeepSearch.

I built this while dealing with large SMB shares over WiFi, where traditional directory scanning tools were either too slow or caused unnecessary network load.

The main idea is a layered scanning strategy: - Minimize SMB round-trips - Use Rayon for controlled parallelism - Avoid spawning excessive IO-bound tasks - Keep memory usage predictable

This is not meant to be a fancy UI tool, but a fast and deterministic building block that can be embedded into other systems.

The project is open-source (Apache 2.0). I'm especially interested in feedback on: - The concurrency model - SMB traversal strategy - Any edge cases you've seen in similar systems

Thanks for taking a look.