This project provides a modern, responsive web UI for Claude Code agents, enabling users to manage and interact with multiple AI-powered coding sessions from their browser. It targets developers and researchers who use Claude Code agents and offers a more accessible and feature-rich interface than the command-line.

How It Works

The UI acts as a server that communicates with Claude Code agents, managing their state and providing a rich interaction layer. It leverages parallel background processing for multiple sessions, allowing users to stream and manage conversations simultaneously. Key features include task management (resume, archive, fork), push notifications, and an integrated dictation feature powered by Gemini 2.5 Flash for precise voice input.

Quick Start & Requirements

• Install/Run: npx cui-server or npm install -g cui-server
• Prerequisites: Node.js >= 20.19.0, logged into Claude Code or valid Anthropic API key. For dictation: Gemini API key (set GOOGLE_API_KEY env var).
• Access: http://localhost:3001/#your-token (token displayed on server start).
• Configuration: Stored in ~/.cui/config.json.
• Docs: Configuration details and usage are described in the README.

Highlighted Details

• Modern, responsive UI for browser access.
• Parallel background agents for simultaneous sessions.
• Task management: resume, fork, archive conversations.
• Gemini 2.5 Flash powered dictation for voice input.
• Push notifications via ntfy for task completion alerts.

Maintenance & Community

The project encourages contributions via issues and stars. A CONTRIBUTING.md file is available for pull requests.

Licensing & Compatibility

The README does not explicitly state a license. Compatibility for commercial use or closed-source linking is not specified.

Limitations & Caveats

Dictation feature requires a Gemini API key and shares audio data with Google. Dictation is only available via HTTPS on iOS. The project does not specify a license, which may impact commercial adoption.