Skip to content

lot-map adds a live map of Claude Code activity

A small TypeScript plugin for Claude Code draws tool calls, model steps, hooks and subagents as pulses on a diagram beside the chat.

Draftship Editorial4 min read

On 2026-11-09 a repository called alexeyw/lot-map appeared on GitHub. By 2026-11-14 it had 759 stars and 87 forks, all gathered in those five days. The project is a plugin for Claude Code, and the idea is narrow: instead of reading a scrolling list of tool calls, you watch a diagram beside the chat light up as the agent works. The lot-map Claude Code plugin is written in TypeScript, released under the MIT license, and described by its author as local only and available in 7 languages.

What the map draws

The readme for alexeyw/lot-map describes a ring of stations around a central core that says CLAUDE by default: You, Model, Memory, Files, Shell, Network, Browser, Agents, Skills and Hooks. An event sends a pulse from the centre to the relevant station, and a second pulse returns when the call ends. Errors and refusals leave a red dot with a count on that station until your next message.

A spoke grows thicker and brighter the more its station is used. A thin route line joins the last few stations the main loop passed through, and each leg stays for 40 seconds before fading over 20. Two clock hands circle a running station, the inner one lapping in a minute and the outer in ten. A ring around the centre tracks the context window, with the percentage above the name, turning amber at 70% and red at 90%. Subagents still at work appear as small dots circling the Agents station.

Below the map sits a line naming what runs now, plus the last 16 rows of activity, marked for running, done, failed or blocked, and a change of model. The clip in the repository is labelled as a made-up session replayed through the mod's own drawing code, so it is an illustration rather than a recording of real work.

Install and requirements

Inside a terminal session you run /plugin install lot-map --marketplace alexeyw/lot-map, answer yes to add the marketplace and pick a scope. From a shell the equivalent is claude plugin install lot-map --marketplace alexeyw/lot-map, with a fallback path in the readme if the plugin is not found. In the desktop app you use the plus button next to the prompt box, then Plugins, then Add marketplace. Updating takes two commands, because the marketplace has to be refreshed first or the second reports the installed version as the latest.

In a terminal the pane needs at least 144 columns to open by itself; below that you type /lot-map, and it draws text only unless you ask for the picture and your terminal supports images. The readme says it was built and tested on Claude Code 2.1.321 on macOS, with users reporting it working on Windows. In the VS Code extension the pane is short, so three buttons there zoom the map across six widths from 160 to 400 pixels or hide it entirely.

Limits and open questions

Classification of shell work goes by the command's name, not by what it does. A script that opens a connection itself, such as python -c, shows as Shell, and npx counts as Network only with -y. Commands wrapped in sudo, env or timeout are read as the command they run. There are many options, including a language setting with 7 codes, theme, a reduced motion mode, text size and a cell shape value for terminal pictures. The source does not mention Linux testing, does not give a version number beyond a changelog, and says nothing about agents other than Claude Code.

What this suggests for coding agents

This reads as a gap builders feel rather than a technical advance. An agent acts through many tools, and the default view of that work is a linear log, which is hard to scan for a stuck loop or an unexpected network call. A map is inference at a glance rather than more capability, and the five-day star count likely reflects how many people run long Claude Code sessions and want to see the shape of them. Watching the map does not change what the agent produces.

For people building apps with AI

The same invisibility applies to any tool that writes code on your behalf. Draftship is an AI app builder where someone describes an app in plain language and it writes the code, builds it and hosts it, with a built-in database and user sign-in. It is not affiliated with alexeyw or lot-map, and nothing in the announcement suggests the plugin works with it.

Primary sources

This article was drafted with AI assistance from the primary sources listed below and checked automatically against them. An editor hasn't reviewed it yet.