Entroly

Name: Entroly
Availability: InStock
Author: juyterman1000

Supports UI

by juyterman1000

Cut AI token costs by 70-95% with a self-evolving codebase compression daemon for LLM context windows.

0 stars

0 views

Works in:CursorCodexWindsurf

Exposes:ToolsResources

View on GitHub Watch demo Docs

What it does

Entroly is a high-performance codebase compression engine that solves the 'blind spot' problem in AI coding tools. It compresses entire repositories into a razor-sharp context window, allowing models like Claude and GPT to see the full project structure without burning through token budgets.

Tools

compress: Compresses any content (code, JSON, logs) to a specific token budget.
compress_messages: Optimizes full LLM conversation histories to retain critical context while removing noise.
wrap: Wraps existing coding agents (Claude Code, Cursor, Codex) to automatically optimize their context delivery.
proxy: Provides an HTTP proxy (localhost:9377) to inject compression into any OpenAI/Anthropic compatible app.

Installation

Add to claude_desktop_config.json:

{
  "mcpServers": {
    "entroly": {
      "command": "pip install entroly && entroly go"
    }
  }
}

Supported hosts

Claude, Cursor, Codex, Aider, GitHub Copilot, Windsurf, Cline, Cody

Quick install

pip install entroly && entroly go

Information

Pricing: free
Published: 5/3/2026
stars

Related Apps

Choose your AI client and follow the steps below.

Codex

Add to AGENTS.md or mcp config entry

Cursor

Add to mcp.json in ~/.cursor/ or project .cursor/

Windsurf

Add to mcp_config.json entry

Claude Desktop

{"mcpServers": {"entroly": {"command": "pip install entroly && entroly go"}}}