OpenAI builds and deploys advanced AI models like GPT-4o for autonomous agents and workflows.
rook
0Popularity
About rook
rook is a local-first, open-source dashboard that discovers, launches, monitors and steers multiple AI coding agents—Claude Code, Codex, Aider and Gemini—from a single interface. It runs entirely on the user’s machine and provides a live roster showing each agent’s status in plain English, highlighting blocked or looping agents before they consume excessive tokens. The tool pulls full PR and issue context from GitHub, including descriptions, commits, linked issues, reviews and comments, and displays it alongside the agent’s diff for review. Users can comment directly on the diff within rook, and those comments are automatically injected into the agent’s session as tracked threads, preserving context without copy-paste. rook also tracks token usage and cost per model and project, with spending windows and trend data, and includes a quality score per run to help assess agent efficiency. Past sessions are preserved and can be resumed with full context, including the same conversation history. The application is installed via a single command and opens a browser automatically, binding to localhost.
Key features
- Live roster of all active agents with status in plain English
- Stuck & loop watchdog to flag agents consuming excessive tokens
- Full PR and issue context pulled live from GitHub
- Diff-first review with in-app commenting and tracked threads
- Per-model and per-project token cost tracking with trend data
- Session history preservation and resume capability
- Keyboard-first navigation and interface
- Single binary installation with automatic browser launch
Use cases
- Monitoring multiple AI coding agents simultaneously for efficiency
- Reviewing and steering agent-generated code via in-app diff comments
- Tracking token usage and cost across different models and projects
Pros
- Local-first execution with no external dependencies
- Supports multiple AI coding agents from one interface
- Preserves full session history for resume capability
- Tracks token usage and cost per model and project
- Keyboard-first navigation and diff-first review workflow
Cons
- Requires local installation and execution
- Limited to supported AI coding agents (Claude Code, Codex, Aider, Gemini)
- No cloud or remote access features mentioned
Frequently asked questions about rook
What is rook and what does it do?
rook is a local-first, open-source dashboard designed to discover, launch, monitor, and steer multiple AI coding agents—such as Claude Code, Codex, Aider, and Gemini—from a single interface. It runs entirely on the user’s machine and provides real-time visibility into each agent’s status.
Who is rook for?
rook is intended for developers and teams who work with multiple AI coding agents and need a centralized way to manage, monitor, and review their outputs. It is particularly useful for those who want to maintain full control over their data and workflows.
How does rook integrate with GitHub?
rook pulls full PR and issue context from GitHub, including descriptions, commits, linked issues, reviews, and comments. This context is displayed alongside the agent’s diff for review, ensuring agents have access to the necessary information.
Can I interact with agents directly within rook?
Yes, users can comment directly on the diff within rook, and those comments are automatically injected into the agent’s session as tracked threads. This preserves context without requiring manual copy-paste.
Does rook track token usage and costs?
rook tracks token usage and cost per model and project, providing spending windows and trend data. This helps users monitor efficiency and manage resources effectively.
How do I get started with rook?
rook can be installed via a single command, and it automatically opens a browser interface bound to localhost. No additional setup is required beyond running the installation command.
