tokless

Save tokens on AI coding agents — no performance loss.

Introduction

A unified CLI to install and update every token-saving plugin for your AI coding agents — fast, efficient, and without hurting how the agent performs.

Supported agents

• Claude Code
• OpenCode
• Codex

Supported tools — each installed from its official source and wired per its own docs:

Tool	What it does
RTK	Shrinks noisy bash/tool output before the model sees it
Caveman	Makes the agent answer in terse, token-light prose
CodeGraph	Lets the agent query a code graph instead of reading whole files
Context-Mode	Runs data-heavy work in a sandbox, returns only what matters

Each tool targets a different source of token waste, so they complement each other with no overlap or conflict.

Install

# macOS / Linux
curl -fsSL https://raw.githubusercontent.com/HoangP8/tokless/main/scripts/install.sh | bash

# Windows (PowerShell)
irm https://raw.githubusercontent.com/HoangP8/tokless/main/scripts/install.ps1 | iex

Commands

tokless              Install + wire everything (default; safe to re-run)
tokless update       Show the version diff and upgrade the four tools
tokless doctor       Show what's wired up; warn about anything broken
tokless uninstall    Remove everything tokless ever touched
tokless self-update  Update the tokless CLI itself

Flags:

--agents <list>   Limit to a subset: claude,opencode,codex
--dry-run         Show what would change without writing anything
--verbose         Show every step

tokless                              # interactive: pick agents, wire all four tools
tokless --agents opencode --dry-run  # preview, no writes
tokless doctor

After running it, restart your agents so they pick up the new config.