siteboon/claudecodeui
Claudecodeui
Claudecodeui is an open-source web interface that orchestrates multiple AI coding agents from different providers—including Claude Code, Cursor CLI, Codex, and Gemini CLI—side by side in isolated cloud environments. It functions as a multi-provider orchestration platform, allowing users to run agents from different tools within the same workspace without being locked into a single vendor.
The platform runs each agent session inside a hypervisor-level Docker sandbox that isolates filesystem, network, and process access, with sessions persisting in the cloud to survive network disconnection or laptop closure. A central controller routes prompts and responses to manage concurrent sessions across multiple CLI backends, while an SSH-tunnelled bridge exposes cloud sessions to local editors for file and terminal access. The system includes a plugin-based extension system that loads custom frontend tabs and backend services, with HTTP and WebSocket calls proxied through the host application.
Users interact with agents through a built-in chat interface, an embedded shell terminal, and a file browser with syntax highlighting. The platform provides Git repository management, session and project management, and team configuration sharing for distributing MCP servers, context files, and environment variables across team members. Agent sessions can be launched from phones, browsers, or APIs and resumed from different devices or IDEs, with background task triggering available for asynchronous completion retrieval.
The application is deployable on personal or team infrastructure via npx or npm, with a managed cloud environment option that provisions persistent Docker sandboxes accessible through SSH, VS Code, or a browser terminal.
Features
- AI Coding Agent Platforms - An open-source web interface that runs Claude Code, Cursor CLI, Codex, and Gemini CLI side by side in isolated cloud environments.
- Plugin-Extensible Agent Interfaces - Extending the agent interface with custom tabs, backend services, and third-party integrations through a plugin system.
- Multi-Provider Orchestrators - Running multiple AI coding agents from different providers side by side in the same workspace without being locked into a single tool.
- Persistent Cloud Workspaces - Keeping AI agent sessions running in isolated cloud containers that survive laptop closure, network loss, or sleep.
- Cloud-Hosted Agent Sandboxes - A platform that provisions persistent Docker sandboxes for AI coding agents accessible via SSH, VS Code, or browser terminal.
- Multi-Agent Orchestration - Manages concurrent AI coding agent sessions by routing prompts and responses through a central controller that coordinates multiple CLI backends.
- Bidirectional Agent Streams - Bidirectional live data channel between frontend and backend that pushes agent output, file changes, and terminal updates without polling.
- Interactive File Editors - Use an interactive file tree with syntax highlighting and live editing capabilities to browse and modify project files.
- AI Agent Session Managers - Provides a chat interface, file explorer, Git integration, session management, and mobile access for agent interactions.
- Plugin-Extensible AI IDEs - An extensible development environment with a plugin system for adding custom tabs, backend services, and third-party integrations.
- Hypervisor-Level Docker Sandboxes - Runs each agent session inside a hypervisor-level Docker sandbox that isolates filesystem, network, and process access.
- Coding Agent Workspaces - An open-source application deployable via npx or npm that runs AI coding agents on personal or team infrastructure.
- Built-in - Access a chat interface, file explorer, Git integration, session management, and mobile support for agent interactions.
- Cloud-Persistent Agent Sessions - Keeps agent environments alive in isolated containers that survive network disconnection by storing session state on the remote host.
- Cloud Session Tunnels - Connect to an existing cloud-based agent session from VS Code or Cursor over SSH, preserving all files and changes.
- Tool Permissions - Sets environment variables, tool permissions, and MCP server connections to tailor the workflow.
- AI Chat Interfaces - Send prompts to supported AI agents and receive their responses through a built-in chat interface.
- Git-Installed Plugin Tabs - Installs plugins from Git repositories that appear as new tabs, rendering custom frontends and optionally running backend servers.
- Asynchronous Agent Task Triggers - Start an agent session from a phone or API and retrieve the completed work later without needing real-time interaction.
- Embedded CLI Terminals - Run CLI commands for the connected AI agent directly from a shell terminal integrated into the user interface.
- Git Repository Managers - View file changes, stage modifications, commit work, and switch branches through a visual interface.
- Frontend-Backend - Loads custom frontend tabs and backend services from a plugin directory, proxying HTTP and WebSocket calls through the host application.
- Configuration Distributors - Distribute MCP servers, context files, environment variables, and tool configurations across team members in isolated containers.
- Vanilla JavaScript Plugin Rendering - Mounts any HTML, charts, or forms into a plugin tab using vanilla JavaScript and receives the current project, session, and theme.
- Self-Hosted AI Infrastructure - Deploying the open-source agent workspace on personal or team infrastructure with no data leaving the network.
- Agent Workspace Deployments - Install and run the user interface on a personal machine or server using npx, npm, or a local development setup.
- Authenticated Plugin Proxies - Proxies authenticated HTTP requests and WebSocket connections from the plugin frontend to its backend server for real-time data streaming.
- Plugin Installation and Management - Builds, installs, and manages plugins that add custom tabs, backend services, and third-party integrations.
- Plugin Backend Servers - Starts a Node.js subprocess per plugin that can access the filesystem, call external APIs, use npm packages, and keep state in memory.
- Cross-Device Session Launches - Launch an AI coding agent session from a phone, browser, or API and resume it later from a different device or IDE.
