FastFileScrape follows the same principles as the FastJava ecosystem:
“No copies. Ever. Critical JNI path. Native‑first performance.”
Even though FastFileScrape v0.1.0 is pure Java, its architecture is designed for a future native backend.
Only essential logic. No streams, no lambdas, no unnecessary abstractions.
Stable ordering, stable chunking, stable output.
The Java version mirrors the structure of the future native version.
JSONL output and chunking optimized for LLM ingestion.
Matches the structure of FastFileIndex, FastFileSearch, FastFileWatch, etc.