JSPM

  • ESM via JSPM
  • ES Module Entrypoint
  • Export Map
  • Keywords
  • License
  • Repository URL
  • TypeScript Types
  • README
  • Created
  • Published
  • Downloads 66
  • Score
    100M100P100Q80214F
  • License MIT

Local web UI for browsing AI coding session history across multiple providers

Package Exports

  • agents-run
  • agents-run/dist/index.js

This package does not declare an exports field, so the exports above have been automatically detected and optimized by JSPM instead. If any package subpath is missing, it is recommended to post an issue to the original package (agents-run) to support the "exports" field. If that is not possible, create a JSPM override to customize the exports field for this package.

Readme

Agents Run

Browse AI coding session history from multiple tools in a unified web UI

npm version License: MIT

Agents Run Demo

The README demo is generated from synthetic sessions, not from local history.


Run the project simply by executing

npx agents-run

Or, after installing globally, use:

agents-run

The browser will open automatically at http://localhost:12001.

Features

  • Multi-provider support - Browse conversations from Claude Code, Codex CLI/Desktop, and Gemini CLI in one place
  • Token usage & cost tracking - Per-session cost breakdown with model-aware pricing for all providers
  • Real-time streaming - Watch conversations update live as Claude responds
  • Search - Find sessions by prompt text or project name
  • Filter by project - Focus on specific projects
  • Resume sessions - Copy the resume command to continue any conversation in your terminal
  • Collapsible sidebar - Maximize your viewing area
  • Dark mode - Easy on the eyes
  • Clean UI - Familiar chat interface with collapsible tool calls

Usage

Install globally via npm:

npm install -g agents-run

If you prefer Homebrew:

brew tap SunZhiC/agents-run
brew install agents-run

Then run it from any directory:

agents-run

The browser will open automatically at http://localhost:12001, showing your AI coding sessions across supported providers.

agents-run [options]

Options:
  -V, --version        Show version number
  -p, --port <number>  Port to listen on (default: 12001)
  -d, --dir <path>     Claude directory (default: ~/.claude)
  --no-open            Do not open browser automatically
  -h, --help           Show help

How It Works

Agents Run reads conversation history from multiple AI coding tools and presents them in a unified web interface:

Provider Data Directory Features
Claude Code ~/.claude/ Full token usage with cost breakdown, session rename/delete, resume
Codex CLI/Desktop ~/.codex/ Token usage with cost breakdown, resume
Gemini CLI ~/.gemini/ Token usage with cost breakdown

The interface includes:

  • Session list - All your conversations across providers, sorted by recency
  • Project filter - Focus on a specific project
  • Conversation view - Full message history with tool calls
  • Token usage & costs - Model-aware pricing (dynamically resolved per session)
  • Session header - Shows conversation title, project name, and timestamp
  • Resume command - Copies the command to resume the conversation
  • Real-time updates - SSE streaming for live conversations

Requirements

  • Node.js 20+
  • At least one of: Claude Code, Codex CLI, or Gemini CLI installed and used

Development

# Clone the repo
git clone https://github.com/SunZhiC/agents-run.git
cd agents-run

# Install dependencies
pnpm install

# Start development servers
pnpm dev

# Build for production
pnpm build

# Refresh the README demo GIF with synthetic data
pnpm demo:gif

Release

# Publish a new npm release
npm publish

# Sync the Homebrew tap formula to the latest published npm version
pnpm sync:homebrew

License

MIT