SuperCompress
Reduce input tokens before LLM calls by ~65% while preserving answer-critical evidence. Open source, CPU-only, ~5K parameters.
Overview
SuperCompress is open-source prompt compression that cuts LLM token costs by ~65%. Reduce input tokens before OpenAI, Claude, Gemini, RAG, chatbot, and agent calls while preserving answer-critical evidence.
SuperCompress is an open-source tool designed to reduce input tokens before LLM calls by approximately 65%, aiming to lower API costs while preserving critical evidence needed for accurate responses.
Who Is It For
Developers and organizations looking to optimize costs associated with LLM API calls, particularly in applications involving chatbots, AI search, and support agents.
Users seeking a comprehensive commercial solution with extensive support or those requiring a GUI-based interface for managing LLM interactions.
Strengths & Weaknesses
Effectively compresses various types of context data, such as chat history and support transcripts, before LLM calls, thereby potentially reducing costs and improving efficiency.
Limited visibility into specific pricing details and the absence of a free trial may hinder adoption for some users. (AI-inferred; may be outdated – founders can correct this)
Alternatives to SuperCompress
Community
No comments yet – be the first to share your experience.