Remote-Code

⚠️ EXPERIMENTAL SOFTWARE - This project is in active development. Expect breaking changes on each commit. No backwards compatibility will be supported yet.

Remote AI agent orchestration platform: Manage multiple coding agents, projects, and tasks from anywhere via web terminals, secure tunnels, and mobile app.

What Remote-Code Does

Remote-Code is a web-based development environment that lets you manage multiple AI coding agents from anywhere. Unlike traditional development setups that tie you to a single machine, Remote-Code gives you a centralized command center that's accessible from any device with a browser.

🎛️ Manage - Complete AI Agent Orchestration

Multi-Agent Coordination

• Run multiple AI coding agents (Claude, Aider, Gemini, etc.) simultaneously
• Each agent operates in isolated tmux sessions to prevent conflicts
• Automatic agent detection and configuration
• Real-time monitoring of all active agents through web terminals

Project & Task Organization

• Organize development work across multiple projects and repositories
• Create and assign specific tasks to different agents
• Track task progress and execution history from a unified dashboard
• Configure base directories with custom setup/teardown automation

Automated Workflow Management

• Automatic git worktree creation for isolated development environments
• Integrated dev server management per project
• Custom setup and teardown commands for different project types
• Session persistence across disconnections

Real-Time Control

• Live terminal access to all running agents via WebSocket connections
• Send commands or guidance to agents mid-execution
• Monitor output and progress in real-time
• Complete visibility into what each agent is working on

🌍 Anywhere - True Location Independence

Cross-Device Access

• Full functionality through any modern web browser
• Works on desktop, laptop, tablet, and mobile devices
• Native mobile app in development for enhanced mobile experience
• No client software installation required

Secure Remote Access

• Cloudflare tunnel integration for secure public access
• ngrok support for quick URL generation
• Self-hosted architecture - your data stays on your infrastructure
• Access your development environment from coffee shops, airports, or home

Persistent Sessions

• tmux-powered sessions continue running even when you disconnect
• Resume work from any device exactly where you left off
• Background task execution - agents keep working while you're away
• Real-time notifications when tasks complete or need attention

How It Works

1. Set up projects with base directories and configuration
2. Configure AI agents (Claude Code, Aider, etc.) through the web interface
3. Create tasks and assign them to specific agents
4. Launch agents in isolated tmux sessions with dedicated git worktrees
5. Monitor and control everything through the web dashboard
6. Access from anywhere using secure tunnel connections

Key Features

• Web-based terminals with full terminal functionality via WebSocket
• tmux session management for persistent, isolated agent environments
• Git worktree automation for conflict-free parallel development
• Project lifecycle management from setup to cleanup
• Editable base directories: update path, git flag, and setup/teardown commands from the project page
• Real-time dashboard showing all projects, tasks, and agent status
• Secure tunnel integration for remote access (Cloudflare, ngrok)
• Mobile-responsive interface with native app in development
• Agent communication - send instructions to running agents
• Background processing - agents work autonomously

Technology Stack

• Backend: Go with WebSocket support, SQLite database
• Frontend: SvelteKit with Tailwind CSS
• Terminal: tmux integration with WebSocket proxying
• Database: SQLite with comprehensive project/task/agent modeling
• Remote Access: Cloudflare tunnels, ngrok integration
• Mobile: Native app in development

Getting Started

# Install dependencies
make install

# Build the application
make build

# Run the server
make run

The server starts on http://localhost:8080 by default.

Use Cases

• Parallel AI Development: Run multiple agents on different features simultaneously
• Remote Code Reviews: Monitor and guide AI agents from your phone or tablet
• Distributed Team Coordination: Share agent sessions and project access
• Long-running Tasks: Start complex refactors, monitor progress remotely
• Mobile Development Oversight: Check build status and agent progress on the go
• Cross-location Continuity: Start work at the office, continue from home

What Makes Remote-Code Different

Remote-Code combines the orchestration power of dedicated agent management tools with the accessibility of web-based development environments. Instead of being limited to a single machine or requiring complex SSH setups, you get a comprehensive platform that scales from solo development to team coordination - all accessible from any device, anywhere.

The focus is on orchestration at scale and access from anywhere - two capabilities that become essential when AI agents handle more of your development workload.

---

Built for the era of AI-assisted development where managing multiple agents remotely is as important as the code they write.