lot-map: A Live Map of What Claude Code Is Doing
A live map beside your Claude Code chat: tool calls, model steps, hooks and skills light up as they happen, with the context window as a ring in the middle.
Project intro
When you work with Claude Code, all you get is a scrolling list of commands. lot-map adds a live map in a pane beside the chat: every tool call, model step, hook and skill sends a pulse from the middle out to its station — You, Model, Memory, Files, Shell, Network, Browser, Agents, Skills and Hooks — so you can see at a glance where the work is going instead of reading line after line.
A thin line joins the last few stations the main loop visited, and spokes grow thicker and brighter the more a station is used, so when the session ends the map shows where the effort went. Two clock hands grow around a station while a call runs; errors and refusals turn red and leave a dot until your next message. The ring around the middle fills as the context window fills — amber at 70%, red at 90% — and when the model changes, the centre rings and the feed notes it, like "opus 5.5 → fable 5.2".
It works in the desktop app, in VS Code with zoom and hide buttons, and in a terminal that can show pictures; a plain terminal still gets the running feed of the last 16 rows. It speaks seven languages, keeps nothing beyond the session except the last model used per folder, and makes no network requests, reads no files and calls no model of its own. It picked up 750 stars and 87 forks in its first five days.
You install it as a Claude Code plugin, and it was built and tested on Claude Code 2.1.321 on macOS, though people report it working on Windows too. In a terminal it needs a wide window — at least 144 columns — or you open the map with /lot-map. The code is MIT licensed.
Actions
Tech stack
Project info
- Project category
- Developer Tools
- Project status
- Online
- Collaboration intent
- None