01. What Oligos Does
Per-Bit Packing
Compressed words are mapped directly to individual bits. Since a byte contains eight bits, a single byte can encode up to eight complete words.
Dimensional Bit Mappings
Every bit references its own dictionary rather than a single global table. Each bit can address up to 265 entries, allowing one byte to reference as many as 2,120 possible words.
02. Reasoning
But.. why?
A major part of Oligos is its ability to sustain a level of compression with virtually any size of data that many mainstream compressors cannot. Oligos is also built for the age of AI, and is heavily architected to deal with the process of machine learning.
This is a huge plus for Stellr amidst its pursuit to achieve effecient Medium Language Models. This enables almost 5x the value for literally 5x less the size of a SOTA LLM.
Read MoreStatic 2-bits-per-byte
Completely randomized streams of words
Expands repetitive compression over time
Adapts unique sequences for higher efficacy
GPU-accelerated compressions
Coming soonDeveloper-owned custom dictionaries
Mitigable metadata (down to few bytes)
0.1 MB total overhead for 550k+ words + metadata
Works anywhere, on virtually anything
03. The Explicits
Single-Bit Packing
Designed for smaller payloads where efficiency is critical. Each byte references up to eight words through strict adherence to an existing dictionary.
Dimensional Bit Mapping
Expands each active bit into its own vocabulary space, allowing a single byte to reference thousands of possible words.
Once a word is identified in the broad dimensional bit mapping, it expands the entire map into a single stream of bytes and locates the exact offset. The receiving end which decodes the message will do the same.
Proofs
Claims are meaningless without verification. These are measurable demonstrations of Oligos compression behavior, performance, and architecture.
Oligos vs ZLIB
Compression benchmark comparing Oligos against traditional deflate-based compression.
Socket Transfer Test
Validating compressed data transmission through real network streams.
Dimensional Compression
Demonstrating multi-dimensional mappings and expanded dictionary references.
Dimensional Compression Walkthrough
Live demonstration of dimensional mappings compressing and resolving encoded data.
Developers
Planned Adaptations
Oligos will begin to be adapted by most of Stellrs' products, SDKs, APIs, and specifications as it continues to mature. Stellr also plans to expand general availability via official APIs in C, C++, Rust, Go, and Python for Orrin Developers.
Multi-Language Official APIs
First-class bindings planned for C, C++, Rust, Go, and Python — systems languages first for zero-overhead performance.
Stellr Ecosystem Integration
Progressive adoption across Terminal Sessions, models, SDKs, and core infrastructure as Oligos matures.
Orrin Developer Priority
Early access and reference implementations prioritized for builders of real-time agents, tools, and high-frequency services.
Minimal. Adaptive. Extreme.
Pioneering real-time stream compression for
high-density networks.