Lossless Compression Python Packages

glyd

Lossless AI compression: 33% less GPU memory, bit for bit.

zxc-compress

Lossless compression C library built for ultra-fast decode. Faster than LZ4 (40-75%+ on ARM64), at a better ratio. Official Rust, Python, Node.js, Go and WASM bindings. Write-once/read-many, with optional seekable archives for O(1) random access.

tomoz

Learned, deterministic lossless compression for medical image volumes: byte-exact DICOM archives, an S3 gateway that compacts PACS storage, Rust, Python and WebAssembly.

audioreconstructor

An adversarial neural network that restores the high fidelity audio lost during lossy compression

neural-weight-compression

Neural Weight Compression: lossless LLM weight compression decoded inside the CUDA matvec. BF16 31 % smaller and faster than cuBLAS; fp8 13 % smaller at native fp8 speed. Bit-identical weights.

bigsmall

Lossless AI model compression - ~34% smaller with bit-identical weights; the autopilot profiles your machine, picks the highest fidelity that runs, and streams models bigger than your RAM.

pyrazobill

Lossless log compaction engine and exact information ledger in C++20 with DuckDB and Python extensions

mlx-dfloat

Run DFloat11 losslessly compressed BF16 checkpoints on Apple Silicon with MLX and a Metal decode kernel; today that runs FLUX.1 on a 32 GB Mac with weights about 30% smaller

densely

Lossless context compression for LLMs: 2x-8x fewer tokens, byte-exact reconstruction (sha256-verified). MCP server for Claude Code & Cursor.

foveance

A lossless token-optimization codec for LLM agents — cut input tokens up to 82%, nothing dropped, every fact kept. Drop-in proxy + one-line Python API for Claude Code, Codex, Cursor, OpenAI, Anthropic & Ollama. pip install foveance

compression-lab

Adaptive lossless compressor with a versioned file format and reproducible benchmark evidence

zpe-iot

ZPE-IoT V0.0: DETERMINISTIC IOT SENSOR CODEC: 1D Time-Series | Chemosense | 17x Mean Compression | Rust Core | PyO3 Native | Edge-Deployable

zpe-xr

Smaller, faster, less accurate VR hand-frame codec. The float16+zlib comparator still wins on fidelity (gate 0/5). ContactPose evaluation lane; PyPI v0.3.1.

mfc-codec

Lossless compression for neuromorphic event-camera data. 47.6× / 1.51 bpe / 0.55 ms per N-MNIST sample. First shipping product of the Mirror Fractal Method. Patent pending. Live demo: codec.mirrorfractal.com

zpe-ink

Python .zpink digital-ink codec — bit-exact roundtrip on UJI, CROHME, DigiLeTs (Hausdorff = 0.0 px). PyPI: zpe-ink v0.1.1.

zpe-neuro

ZPE-Neuro V0.0: DETERMINISTIC NEURAL SIGNAL CODEC: Extracellular Spikes | Neuropixels | DANDI-Anchored | IBL Validated | Spike Train Transport

zpe-mocap

ZPE-Mocap V0.0: DETERMINISTIC MOTION CAPTURE CODEC: Joint Angles | Positional Fidelity | Search Ranking | Sub-ms Query | Skeletal Transport

zpe-music

ZPE-Music V0.0: DETERMINISTIC SYMBOLIC SCORE CODEC: MusicXML Roundtrip | Note-Local Expression Refinement | Attack | Release | Dynamics

zpe-smell

ZPE-Smell V0.0: Surrogate-receptor odor codec on a fixed 35-episode benchmark | Spearman 0.9996 | 0% mixture collision | Empirical olfactory perception out of scope

zpe-touch

ZPE-Touch V0.0: DETERMINISTIC TOUCH STREAM CODEC: Contact Geometry | Thermal | Vibrotactile | Proprioceptive Branch Isolation | Rust Core

zpe-mental

Tunnel, spiral, lattice, cobweb — byte-exact codec on the bounded four-form corpus. Rust-native fast path with a Python reference. Not a cognition or clinical claim.