Problem solved

Managing multiple command-line AI agents simultaneously is challenging, especially when accessing them remotely or keeping sessions alive across disconnections. tlbx solves this by providing a persistent, browser-accessible workspace where each agent runs in its own PTY session, enabling parallel execution and easy switching without requiring complex terminal multiplexers or SSH setups.

Use cases

• Running multiple AI coding assistants (Claude Code, Codex, OpenCode) in parallel for accelerated development
• Remotely monitoring and interacting with CLI agents from a phone or tablet
• Long-running autonomous code generation or refactoring tasks that require persistent sessions
• Private, self-hosted AI development environment for security-sensitive projects
• Team collaboration on agent-driven coding tasks via shared browser workspace

Audience

• Developers
• ML Engineers
• Data Scientists
• Researchers
• Startups

Pros

• Self-hosted ensures data privacy and control
• Supports multiple CLI agents running in parallel
• Real PTY sessions persist across browser closures
• Accessible from any browser or mobile device
• Open-source and free to use

Cons

• Requires manual setup and configuration on your own server
• Limited to CLI-based agents; no GUI automation
• May consume significant resources when running many agents
• Lacks built-in advanced orchestration or workflow management
• Relatively new project with potentially limited community support