The original title is: "SuperCompress is now on PyPI! pip install supercompress in 1 line"

Short summary

SuperCompress is a lightweight Python library (~5K parameters) that compresses LLM context by 65% while maintaining 100% accuracy on critical information. It uses a CPU-based relevance scorer to eliminate irrelevant lines before sending to the LLM, reducing token costs with ~60ms latency and zero GPU requirement. The tool is now available on PyPI under MIT license with open-source code, an interactive demo, and technical deep-dive.

• •65% token compression with 100% oracle recall on critical information
• •~5K CPU-only model; ~60ms latency; no GPU needed; MIT open-source license
• •Interactive comparison tool and technical deep-dive available; installable via pip

Generated with AI, which can make mistakes.