Idle Compaction in Claude Code
Small update for Claude Code users in VS Code and the Desktop app: long idle conversations get summarized so the next message starts cheaper.
Details
- When a conversation of 200K+ tokens sits idle, Claude Code summarizes it in the background.
- If the prompt cache has expired by the time you send your next message, Claude uses that summary instead of the full history.
- Works in the VS Code extension and the Code tab in Claude Desktop, including for subscribers with telemetry turned off.
- Also runs on Amazon Bedrock, Google Vertex, Microsoft Foundry, and LLM gateways once the one-hour prompt cache is on.
- Shipped in Claude Code 2.1.314. Most users will not notice it beyond faster, cheaper restarts.
Best for
Picking a long coding session back up after a break without paying to reload the whole conversation.
When to use it
It runs on its own. You notice it when you return to a very long conversation after the cache has expired.
Who benefits
People who leave big Claude Code sessions open in VS Code or the Desktop Code tab and come back hours later.
Join the newsletter to get ahead with AI.
I share strategies, new modules, free guides, and tools to get you better at using AI in your work.
Leading a team? See team programs →