Chisle

Cut your AI coding agent's token bill on three axes: terse prose, YAGNI-first code, and tool-output compression. Claude Code, Pi, Cursor, Codex, Gemini + 4 more

What is Chisle?

Cut your AI coding agent's token bill on three axes: terse prose, YAGNI-first code, and tool-output compression. Claude Code, Pi, Cursor, Codex, Gemini + 4 more. Zero deps, published benchmarks including the runs it loses.

Overview of Chisle

What it is

Chisle is a plugin for AI coding agents that reduces token consumption without sacrificing accuracy. It works by injecting a ruleset that makes the agent produce terser prose, write only the code that is needed (YAGNI), and compress tool output before it re-enters the context window. It supports Claude Code, Pi, Cursor, Codex, Gemini, and nine other agents. The project publishes raw transcripts of its benchmark runs, including the ones where it lost. It has zero dependencies.

Key features

  • Injects a zero-fluff ruleset per session to reduce filler, hedging, and manufactured structure in agent output.
  • Applies a YAGNI efficiency ladder: the agent checks whether code is needed, already exists, can be done with stdlib, platform features, or existing dependencies, or can be one line, before writing the minimum code that works.
  • Compresses oversized tool output (e.g., bash, grep, web results) by about 46% via a PostToolUse hook before it re-enters the context window, where it would be billed again on every later request.
  • Extends tool-output compression to MCP servers in any agent via npx chisle mcp wrap.
  • Auto-detects installed agents and wires each one with a single command; --uninstall reverses the changes.
  • Published benchmarks show Chisle produces the leanest answers compared to bare models, caveman, and ponytail rulesets, while maintaining accuracy (156/156 vs 155/156 on auto-graded maths and coding).

Use cases

  • Reducing token costs when running Claude Code, Pi, Cursor, Codex, Gemini, Copilot, OpenCode, or other supported agents.
  • Shrinking context window usage in long agent sessions where tool output is repeatedly billed.
  • Enforcing concise, YAGNI-first coding style across multiple agents with a single ruleset.

Getting started

bash npx chisle bash # or via curl curl -fsSL https://raw.githubusercontent.com/JayPokale/Chisle/main/install.sh | bash powershell # Windows irm https://raw.githubusercontent.com/JayPokale/Chisle/main/install.ps1 | iex Preview first with npx chisle --dry-run, scope with --only claude or --only pi, see everything with npx chisle --help. Remove with npx chisle --uninstall.

License and pricing

The project is open source under the MIT license. No pricing information is stated beyond the open source license.

Based on the project's public README and repository metadata.

Add this badge to your README

Featured on YunoDB Badge styles & help

Maintain this project? Show that it is featured on YunoDB.

Markdown[![Featured on YunoDB](https://yunodb.com/badge/jaypokalechisle.svg)](https://yunodb.com/tools/jaypokalechisle?utm_source=github&utm_medium=badge)
HTML<a href="https://yunodb.com/tools/jaypokalechisle?utm_source=github&utm_medium=badge"><img src="https://yunodb.com/badge/jaypokalechisle.svg" alt="Featured on YunoDB" height="20"></a>