Agent Inbox UI
Web app for human-in-the-loop interactions with AI agents in LangGraph applications
🛡️ AgentReady threat assessment
MAESTRO 7-layer threat model + OWASP AIVSS risk score for Agent Inbox UI, derived from its capabilities.
These scores are auto-generated from public information (the agent's own listing, docs, and repository) using the canonical OWASP AIVSS formula and the MAESTRO framework — an estimate for guidance, not a penetration test, audit, or certification. See the scoring methodology — every score is re-derived by the same automated method as an agent's public evidence changes.
Overview
Agent Inbox is an open-source web application that facilitates human-in-the-loop interactions with AI agents in LangGraph apps. It provides a user-friendly interface for both technical and non-technical users to engage with AI workflows. The tool supports various interaction types, including accepting, editing, or responding to agent actions, and can be integrated with any LangGraph application running locally or deployed.
Key features and capabilities
- Compatible with any LangGraph app (Python and JavaScript/TypeScript)
- Supports multiple human-in-the-loop action types (accept, edit, respond, ignore)
- Customizable interrupt configurations
- Local storage for API keys and inbox settings
- Markdown support for detailed interrupt descriptions
- Easy setup and configuration process
Use cases
- Reviewing and approving AI-generated content
- Editing AI tool calls or arguments
- Providing custom responses to AI queries
- Monitoring and managing multiple AI agent workflows
- Facilitating collaboration between AI systems and human experts