msdickinson

AI Timeline — AI skill for Claude Code

AI community

See what AI coding agents actually do, across one session or thousands.

How to install AI Timeline

This entry records only its repository, not the path inside it, so there is no exact command to give. Open msdickinson/ai-timeline and copy the folder into ~/.claude/skills/, or the file into ~/.claude/agents/.

What AI Timeline does

See what AI coding agents actually do, across one session or thousands. Timelines, tool calls, tokens and cost, plus patterns across whole benchmark runs, compared by model. Reads Claude Code, Cursor, Aider, Cline, SWE-Agent, OpenHands and more; add a small parser for your own setup. Runs in your browser; nothing is uploaded.

Alternatives in AI

  • Codesight — Universal AI context generator 979 ★
  • Nativeprompt — Rewrite your prompt into the native dialect of the model you're actually running (Claude Code / Codex) — by ea 113 ★
  • Sidekick Agent Hub — See what your AI coding agent is doing 80 ★

README

AI Timeline

[![CI](https://github.com/msdickinson/ai-timeline/actions/workflows/ci.yml/badge.svg)](https://github.com/msdickinson/ai-timeline/actions/workflows/ci.yml) [![License: MIT](https://img.shields.io/badge/license-MIT-blue.svg)](LICENSE) [![Web app](https://img.shields.io/badge/web%20app-try%20it-4f8ff7)](https://msdickinson.github.io/ai-timeline/) [![VS Code](https://img.shields.io/badge/VS%20Code-extension-007acc?logo=visual-studio-code&logoColor=white)](#vs-code-extension)

**See what AI coding agents actually do, across one session or a whole benchmark run.**

AI Timeline dashboard showing the built-in sample: a Claude Code session with a team of subagents

Every time you use Claude Code, Cursor, Aider, or Copilot — it saves a full record of everything that happened. Every tool call, every token spent, every file it read, every error it hit. It's all sitting in a folder on your machine right now. You've probably never looked at it.

AI Timeline reads those files and shows you what your AI actually did — timelines, token breakdowns, patterns, and things you'd never see in the chat UI. Load one session, or a whole folder from a benchmark run and look for patterns across all of them, compared by model. Using your own agent setup? Add one small parser file (see [Add your own agent](#add-your-own-agent)).

60 seconds. No install.

  1. Open the web app (or click "load a sample session" — no files needed)
  2. Drop a session file from Claude Code, Cursor, Aider, or any of the supported tools
  3. See what happened — a timeline of every tool call, auto-generated insights, and token breakdowns

Everything runs in your browser. Your data never leaves your machine.

Why use this

  1. See where time went. Your session took 15 minutes — was it thinking, running tools, or stuc