Hystax Releases Open-Source Token Compressor Cutting AI Costs

SUNNYVALE, Calif., Oct. 21, 2026 /CourierPR/ -- Hystax, a cloud infrastructure and AI governance software company, has released OptScale Kompressor, an open-source token compressor that significantly reduces the costs associated with AI coding agents. The tool, which is now available on GitHub under the MIT license, has been shown to cut provider bills by 22.7% without affecting the success rate of the tasks being solved.

OptScale Kompressor works by shortening the build logs, compiler messages, and other machine-generated output that AI coding agents send to a language model. This is achieved without altering the agent or its prompts, ensuring that the model receives all necessary information. According to Hystax, the compressor operates on a rule-based system, making it deterministic and recoverable. It is designed to be format-aware, meaning it can handle different types of output such as logs, compiler messages, JSON, diffs, or source code, each with its own specialized compressor.

In benchmark tests conducted on Terminal-Bench 2.1, a public benchmark of 89 command-line tasks, the compressor demonstrated significant cost savings. Compared to no compression, it reduced the provider bill by 22.7%, which is approximately twice the savings achieved by Headroom, an open-source compressor. The savings were particularly pronounced in log-heavy work, where bills fell by 26% in machine learning training and evaluation tasks and 28% when building software from source.

OptScale Kompressor is available as a single Rust binary for Linux, macOS, and containers, and processes one kilobyte of output in about 0.2 milliseconds. It does not make any outside network calls, ensuring that it operates independently and securely. The tool is designed to be fail-safe, meaning that if compression fails or does not shrink the output, the original data passes through unchanged.

The release of OptScale Kompressor is part of Hystax's broader strategy to optimize cloud costs and enhance AI governance. The company's OptScale AI platform, launched in May 2026, is designed to tie AI spend to delivered work, cut costs through smart routing and token compression, screen every prompt, and keep AI agents within their limits. The full benchmark report, with per-task results and methodology, is available on the OptScale AI blog.

Hystax's commitment to open-source development is evident in the release of OptScale Kompressor, which is expected to benefit a wide range of organizations looking to optimize their AI operations and reduce costs.