ContextCoach

Tracking since May 11, 2026.

Changelog

• Jun 5, 2026
  description

  ContextCoach helps you reduce the processing cost and energy use of your AI use. 
  
  Every time you send a message to Claude.ai or ChatGPT, the AI re-reads the entire conversation from the beginning before answering. The longer the chat — and the more documents you've loaded — the more it has to process. This is called context, and it drives cost and energy use with every single query.
  
  It's hard to tell when the context has grown too large. ContextCoach shows you.
  
  As you chat on Claude.ai or ChatGPT, ContextCoach tracks what's driving your context load: conversation length, uploaded documents, and project files. A color-coded banner shows your estimated token count — tokens are roughly one per word, but uploaded documents can add thousands.
  
  Green — you're in good shape.
  Yellow — context is growing. Consider starting fresh if the topic has shifted.
  Red — context is large. Starting a new chat will cut processing cost and energy use. ContextCoach tells you by how much.
  
  The thresholds are based on what you can control — conversation length and your uploads. Claude.ai's system automatically loads ~15,000 tokens of its own overhead before you type a word (and ChatGPT loads ~4,000). Project knowledge files add more on top. ContextCoach subtracts those fixed costs so the color reflects only what starting a new chat would actually save. That's why a chat showing 30,000 total tokens can still be green if most of it is project files.
  
  When it's time to start fresh, one click on Summarize Prompt asks the AI to summarize your conversation — then one more click opens a new chat with the summary ready to go.
  
  What ContextCoach tracks:
  - Conversation length
  - Uploaded documents and PDFs
  - Project files (on Claude.ai)
  
  Free. No API key. No account. No backend.
  
  ContextCoach runs entirely in your browser. Nothing leaves your computer — your conversations are never transmitted anywhere.
  
  Works on Claude.ai and ChatGPT.

  ContextCoach helps you reduce the processing cost and energy use of your AI use. 
  
  Every time you send a message to Claude.ai or ChatGPT, the AI re-reads the entire conversation from the beginning before answering. The longer the chat — and the more documents you've loaded — the more it has to process. This is called context, and it drives cost and energy use with every single query.
  
  It's hard to tell when the context has grown too large. ContextCoach shows you.
  
  As you chat on Claude.ai or ChatGPT, ContextCoach tracks what's driving your context load: conversation length, uploaded documents, and project files. A color-coded banner shows your estimated token count — tokens are roughly one per word, but uploaded documents can add thousands.
  
  Green — you're in good shape.
  Yellow — context is growing. Consider starting fresh if the topic has shifted.
  Red — context is large. Starting a new chat will cut processing cost and energy use. ContextCoach tells you by how much.
  
  The thresholds are based on what you can control — conversation length and your uploads. Claude.ai's system automatically loads ~15,000 tokens of its own overhead before you type a word (and ChatGPT loads ~4,000). Project knowledge files add more on top. ContextCoach subtracts those fixed costs so the color reflects only what starting a new chat would actually save. That's why a chat showing 30,000 total tokens can still be green if most of it is project files.
  
  When it's time to start fresh, one click on Summarize Prompt asks the AI to summarize your conversation — then one more click opens a new chat with the summary ready to go. 
  
  Working inside a Claude Project or ChatGPT custom GPT? ContextCoach preserves that context too — the new chat opens inside the same project or GPT, so your knowledge files and instructions stay attached.
  
  What ContextCoach tracks:
  - Conversation length
  - Uploaded documents and PDFs
  - Project files (on Claude.ai)
  
  Free. No API key. No account. No backend.
  
  ContextCoach runs entirely in your browser. Nothing leaves your computer — your conversations are never transmitted anywhere.
  
  Works on Claude.ai and ChatGPT.

• Jun 5, 2026
  permissions

  activeTab, storage, tabs

  activeTab, storage

Permissions & access

Permissions

activeTabstorage

Host access

https://claude.ai/*, https://chatgpt.com/*

Screenshots