Rover, the AI background agent for enterprise
Autonomous AI agents that work in the background on your projects. Built for professional startups and engineering teams.
Created by Adhyaay Karnwal.
Rover is an enterprise-grade background agent platform that enables engineering teams to deploy AI agents capable of autonomous software development, testing, and deployment. Inspired by Ramp Inspect, Rover provides isolated, containerized environments where AI agents can safely execute complex tasks.
- Multi-Session Orchestration: Run multiple AI agents simultaneously on different projects
- Full Browser Automation: Agents can navigate, interact, screenshot, and record web applications
- Containerized Workspaces: Each session runs in isolated Docker containers with persistent storage
- Real-Time Collaboration: Observe agent progress in real-time via WebSocket streaming
- Dynamic Routing: Access deployed applications via auto-generated subdomains
- Enterprise Integrations: GitHub OAuth, iMessage notifications, and team collaboration features
- Microservices Architecture: Horizontally scalable services built with Bun and TypeScript
- Browser Isolation: Each session gets dedicated Chrome instances via CDP
- State Management: Redis for session state, PostgreSQL for persistence
- Object Storage: S3-compatible storage (RustFS) for screenshots and recordings
- Reverse Proxy: Dynamic subdomain routing to session containers
- Bun 1.3.8 or later: Install Bun
- Docker and Docker Compose
- Git
# Clone the repository
git clone <repository-url>
cd rover
# Install dependencies
bun install
# Configure environment
cp .env.example .env
# Edit .env with your API keys and configuration
# Start development environment
bun run devOnce running, access the platform at:
| Service | URL | Description |
|---|---|---|
| Web UI | http://localhost:3000 | Main dashboard and orchestrator |
| API | http://localhost:4000 | REST API and WebSocket endpoints |
| OpenCode | http://localhost:3002 | Agent CLI interface |
| Storage Console | http://localhost:9001 | Object storage management |
- Define Projects: Configure development environments with Docker images and system prompts
- Submit Requests: Send tasks to the orchestrator via web UI or messaging platforms
- Agent Execution: AI agents autonomously work in isolated containers
- Real-Time Monitoring: Observe progress via browser streams, logs, and screenshots
- Review & Deploy: Preview changes on generated subdomains and merge when ready
- Getting Started - Complete setup and first project guide
- Architecture - System design and component overview
- Configuration - Environment variables and settings
- Projects - Creating and managing projects
- Sessions - Working with agent sessions
- Browser Automation - Browser tools and capabilities
- Contributing - Development guidelines and contribution process
- Security - Security best practices and considerations
- Troubleshooting - Common issues and solutions
- API Reference - REST API documentation
The Rover interface provides a comprehensive view of active sessions, real-time browser streaming, container logs, and file management—all in a unified dashboard.
This screenshot does NOT capture everything about rover. I plan on adding more soon, but feel free to self host and see for yourself.
- Parallel Development: Run multiple feature branches simultaneously
- Automated Testing: Agents can write and execute test suites
- Code Review: Automated code analysis and suggestions
- Documentation: Auto-generate documentation from code changes
- Rapid Prototyping: Quickly iterate on product ideas
- 24/7 Operations: Agents work continuously on background tasks
- Cost Efficiency: Shared infrastructure for team workloads
- Audit Trail: Complete history of all agent activities
- Access Control: Role-based permissions and GitHub SSO
- Audit Logging: Comprehensive activity logs
- Resource Limits: Container resource quotas and monitoring
- Compliance: Data retention and privacy controls
- Runtime: Bun 1.3.8, TypeScript 5.x
- Frontend: Next.js 16, React 19, Tailwind CSS 4
- Backend: Microservices with Bun HTTP servers
- Database: PostgreSQL 17 with Drizzle ORM
- Cache: Redis 7 for session state
- Storage: RustFS (S3-compatible object storage)
- Browser: Playwright/Chromium via agent-browser
- AI: OpenCode SDK, Vercel AI SDK
AGPL-3.0 - See LICENSE for details.
For issues, feature requests, or questions:
- Review the documentation
- Check troubleshooting guide
- File an issue in the repository
Rover - Enterprise AI agents that actually deliver.