KHENOR / UNIVERSAL MATH LAB / v0.21Any file.
Reconstructed exactly.
A mathematical research engine for arbitrary bytes—not only pictures. Create a .khen archive from any file type, restore the original bytes, and see exactly how much space is saved. Processes files in small blocks, privately on your device.
X = (b₀,b₁,…,bₙ) Xⱼ = Fⱼ(i,θⱼ) ⊕ Rⱼ read(offset, length) → original bytes SHA-256 → verified
Exact GF(2) recurrence discovery • 256 KiB blocks • experimental
KHENOR / UNIVERSAL WORKSPACE
Store mathematical rules. Recover original bytes.
The public KH4.1 codec learns short XOR recurrence equations from byte sequences, alongside the previous exact formula library and Deflate fallback. Advanced sparse-relation synthesis remains private research. This public XOR method uses established linear algebra, not a new mathematical theorem. Neither codec is proven superior to industrial compression.
◉ Processed locally in your browser
01 / ENCODE
Create a .khen file
Choose any file, including ZIP archives, documents, source code, videos, images or binaries. All file formats are accepted, including ZIP, PDF, video and programs. Files stay on your device. Larger files require a browser with direct-to-disk saving.
No file selected.
No server upload: KHENOR reads your selected file locally only when you press Encode.
02 / RESULTS
Actual bytes. No estimates.
Encoded size includes mathematical parameters, block metadata and integrity digests.
No result yet.
Already compressed or encrypted files often cannot shrink further. Smaller than raw bytes does not imply better than Zstandard or other specialized codecs.
03 / DECODE
Restore the original file
Open a KH4.0 or KH4.1 .khen archive; every decoded block is verified before the file is accepted.
No archive selected.
Download a tiny 200-byte .khen archive, then open it under Restore. It recreates a synthetic 512 KiB binary sequence exactly using learned XOR equations. This is not a benchmark on real-world files.
New archives use KH4.1; this decoder also reads previous KH4.0 files. Older decoders cannot open new archives until updated. Advanced KH7–KH9 formats are separate research formats.
We are prototyping a read(offset,length) interface for a future virtual filesystem. A native KHENOR Drive extension is not installed or available yet.
RESEARCH ROADMAP
Beyond compression: access without extraction.
A future KHENOR Drive could expose a compressed document through operating-system file APIs as its original file type. Our research decoder can already read selected byte ranges; a production filesystem, server-scale performance and multi-user security are still future work.
Chromium-based desktop browsers can stream to user-selected disk files; Firefox and unsupported browsers are currently limited to 32 MiB in-memory per operation. Do not treat this prototype as certified for unlimited size. No accounts, commercial quotas, AI reconstruction, encryption or data-center service are active.