TokenEconomizer

TokenEconomizer is a plugin that cuts the running cost of coding agents by learning from past sessions in a repository, for developers.

What it is good for

  • Reducing the cost of running a coding agent on the same project
  • Stopping a coding agent from rediscovering the same project facts
  • Cutting wasted tokens from unused tools loaded in every request
  • Catching a coding agent stuck repeating the same failures

At a glance

Who it is for
Setup difficulty
Technical ยท Guided command-line setup

Level 4 of 5. You copy and paste commands into a terminal by following the instructions. Comfortable territory for vibe coders.

Cost
Free and open source
Type
Skill or plugin
Tasks
Works with
Claude Code, Codex, GitHub Copilot CLI, Pi, opencode, Cursor Agent

How to get started with TokenEconomizer

  1. Check that you have a supported agent signed in and Python 3.10 or newer on your PATH.
  2. Clone the repository with git clone https://github.com/microsoft/TokenEconomizer.git in a folder you will keep.
  3. Add the clone as a plugin marketplace and install the plugin using your agent's plugin commands, such as the Claude Code ones in the README.
  4. Start the agent in one of your repositories and trust the folder when asked.
  5. Run the status command, such as /tokeneconomizer:status in Claude Code, to confirm it works.

From the project's own documentation. If a step does not match what you see, the official pages have the latest instructions.

Before you try it

  • Requires Python 3.10 or newer and one supported coding agent already signed in.
  • Opencode and Cursor Agent support is experimental.
  • Its own model calls run through your agent's plan or login, and their cost is counted in reports.
  • Savings come from learning across sessions, so the first sessions in a repository start with nothing learned.

Momentum

Rank this week
#33
Best rank
#33 week 46, 2026
Stars gained last week
+1,460
GitHub stars
6,846

GitHub numbers as of Nov 16, 2026. Updated weekly.

2,036 4,619 7,202 Oct 31, 2026 Nov 16, 2026
GitHub stars for TokenEconomizer: 2,392 on Oct 31, 2026, 6,846 on Nov 16, 2026. Source: GitHub.

Ranking history

First ranked in week 45, 2026.

Questions

Is TokenEconomizer free?
Yes. It is free and open source (MIT license).
Do I need to know how to code to use TokenEconomizer?
You do not need to write code, but setup means copying and pasting commands into a terminal by following its instructions.
What does TokenEconomizer work with?
According to its own documentation, it works with Claude Code, Codex, GitHub Copilot CLI, Pi, opencode, and Cursor Agent.
How do I get started with TokenEconomizer?
1. Check that you have a supported agent signed in and Python 3.10 or newer on your PATH. 2. Clone the repository with git clone https://github.com/microsoft/TokenEconomizer.git in a folder you will keep. 3. Add the clone as a plugin marketplace and install the plugin using your agent's plugin commands, such as the Claude Code ones in the README. 4. Start the agent in one of your repositories and trust the folder when asked. 5. Run the status command, such as /tokeneconomizer:status in Claude Code, to confirm it works.
What should I know before trying TokenEconomizer?
Requires Python 3.10 or newer and one supported coding agent already signed in. Opencode and Cursor Agent support is experimental. Its own model calls run through your agent's plan or login, and their cost is counted in reports. Savings come from learning across sessions, so the first sessions in a repository start with nothing learned.
What is TokenEconomizer good for?
Reducing the cost of running a coding agent on the same project, stopping a coding agent from rediscovering the same project facts, cutting wasted tokens from unused tools loaded in every request, and catching a coding agent stuck repeating the same failures.

Similar agents to TokenEconomizer

  • caveman

    Cuts token use by making the agent answer tersely and shrinking input text, rather than learning from past sessions.

  • ECC

    Adds planning, testing, review, and memory routines to coding agents, with no focus on cost reduction.

  • CC Safety Net

    Blocks destructive commands and secret file access to keep the agent safe, rather than lowering cost.

  • GSD Core

    Adds a phased planning and verification workflow to coding agents instead of reducing running cost.

  • skills

    skills is a set of skill files that helps coding agents understand what you want, write tested code, and debug, for software developers.

  • Superpowers

    Superpowers is a skill plugin that makes coding agents plan, test, and review their work step by step for people building software with AI.

Last updated Nov 17, 2026. Tracked since Oct 31, 2026. Spotted an error? Tell us.