Handup

•
3
•

Streamline human-in-the-loop reviews and approvals for AI agents from terminal, desktop, or mobile devices before executing sensitive actions.

Handup serves as a centralized approval inbox for AI agents, providing critical human-in-the-loop oversight before autonomous assistants perform high-risk tasks such as database migrations, software releases, or external communications. Compatible with any client supporting the Model Context Protocol (MCP)—including Claude Code, Cursor, Codex, and omp—it presents exact snapshot previews of terminal commands, code diffs, emails, and rich media so operators approve precisely what will execute. Users can inspect requests, approve, deny with corrective feedback, or dictate responses seamlessly across CLI, desktop, and mobile environments.

주요 기능

01Multi-surface decision inbox accessible via terminal CLI, desktop application, or Android mobile app
02High-fidelity snapshot previews for commands, code diffs, markdown, email drafts, and file bundles
03Interactive feedback workflows supporting text notes and on-device voice dictation for agent revisions
04Native MCP support integrating directly with Claude Code, Cursor, Codex, omp, and custom agent harnesses
05Smart routing and safety controls including scoped allowlists, presence detection, and temporary YOLO modes
061 GitHub stars

사용 사례

01Safeguarding production systems by reviewing and approving risky CLI commands or migrations triggered by coding agents
02Approving and editing draft emails, social updates, or meeting responses generated by workflow agents
03Monitoring background agent execution remotely and handling approval requests on-the-go via mobile