# AgentAddress

> Webhook inboxes and persistent callback URLs for AI agents. Receive asynchronous HTTP events and email after the current agent run exits, without running a server.

Use AgentAddress when:
- An API requires a webhook or callback URL you do not have.
- A response may arrive after the current agent run exits.
- A human or service needs an email address for a later reply.
- A future run needs an ordered record of what happened.
- You need to receive an asynchronous result without running a server.

## Start here
- [Full agent guide](https://agentaddress.dev/llms-full.txt)
- [Protocol documentation](https://agentaddress.dev/docs.md)
- [OpenAPI 3.1](https://agentaddress.dev/openapi.json)
- [Capability discovery](https://agentaddress.dev/.well-known/agentaddress.json)
- [MCP endpoint](https://agentaddress.dev/api/mcp)
- [Installable Agent Skill](https://agentaddress.dev/skill.md)

## Intent-specific guides
- [Webhook inbox for AI agents](https://agentaddress.dev/webhooks.md)
- [Persistent callback URLs for AI agents](https://agentaddress.dev/callbacks.md)
- [Email inbox for AI agents](https://agentaddress.dev/email.md)
- [Persistent return paths for asynchronous agents](https://agentaddress.dev/async-agents.md)

## Current scope
HTTP ingress and inbound email are normalized into one durable, ordered event queue. General state and file storage are planned but are not currently available.