A Smarter
Archive Format
Research project: a next-generation archive format with adaptive compression, full-text search, versioning, and encryption. Designed for collections where content matters — not just files. Codename: .apx.
This is a research project. No release date, no public build. Architecture is being proven before any consumer product is built. Listed here for transparency on what we are working on and why.
Why Build a New Archive Format?
Zip and tar were designed in a different era. They treat an archive as a flat list of files — no deduplication, no search, no relationships, no adaptive compression. Every repack is a full copy.
For structured creative and technical projects — codebases, design assets, research data — the archive format itself should understand the content.
Proving the Architecture First
Each milestone must be benchmarked and proven before the next begins. No premature optimization.
SQLite metadata schema, archive inspector, verification, and benchmark framework.
alpha create and alpha extract. Single-pass flat archive, no deduplication yet. Baseline benchmarks established.
Chunk boundaries, rolling hash, content-addressed chunk store. Measure space savings on real project trees.
Per-chunk algorithm selection. Benchmark zstd, brotli, and lz4 across content types. Choose per file signature.
Full-text metadata search, delta versioning, first-class encryption, and migration of the core to Rust for performance.
What It Could Become
If the research proves out, these are the product directions that follow.
Follow the Research
Interested in this problem space? Get in touch.
contact@lpagesapplabs.com