elirantutia/vibeyard

What it solves

Running AI coding agents (like Claude Code or Gemini CLI) in a standard terminal often leads to a disorganized workspace. Vibeyard provides a dedicated IDE-like environment to manage multiple agent sessions, track costs, and organize tasks without juggling multiple terminal windows.

How it works

It acts as a management layer over existing AI CLI tools. It provides a graphical user interface that wraps around PTYs (pseudo-terminals) for tools like Claude Code, Codex CLI, and Gemini CLI. It includes a project dashboard with widgets, a Kanban board for task tracking, and an embedded browser for sending DOM elements as context to the AI.

Who it’s for

Developers who use AI coding agents via the command line and want a more structured, visual way to manage their agentic workflows, session history, and team collaboration.

Highlights

• Multi-session management: Run multiple agents in parallel with a "swarm mode" grid view and isolated configuration profiles.
• Kanban task board: Plan work with cards that can spawn or resume CLI sessions in one click.
• Cost and context tracking: Real-time monitoring of spend, token usage, and context window size per session.
• P2P session sharing: Share live terminal sessions with teammates via encrypted WebRTC connections.
• Embedded browser: Capture DOM selectors and page content to send as precise editing instructions to the AI.
• AI Readiness Score: Evaluates how prepared a project is for AI-assisted coding with suggested fixes.