> ## Documentation Index
> Fetch the complete documentation index at: https://agentmail.manicule.dev/llms.txt
> Use this file to discover all available pages before exploring further.
> For every page in one file, fetch https://agentmail.manicule.dev/llms-full.txt (large).
> To search and read these docs as tools, connect the MCP server at https://agentmail.manicule.dev/api/mcp.
> Every agent surface: https://agentmail.manicule.dev/.well-known/llms-manifest.json

# AgentMail (/)

<!-- agent-signals: reading_time_min: 1 · est_tokens: 305 · updated: 2026-09-29 -->
Related: [Introduction](/introduction.md), [Quickstart](/quickstart.md), [Architecture](/architecture.md)

# Give your agent its own email inbox

> Email inboxes for AI agents. Create inboxes, receive mail, trigger workflows, and reply through one API.

AgentMail provides email inboxes for AI agents: a real address an agent can hand out, watch, and act on. Create inboxes, receive mail, trigger workflows, and reply through one API.

Not sure what to build? Describe what you are working on at [agentmail.to/build](https://agentmail.to/build) and get a blueprint back, with a prompt your coding agent can build from.

## Related

* [/quickstart](/quickstart) - get an API key, create an inbox, and run the full send, receive, and reply loop.
* [/introduction](/introduction) - what AgentMail is and why an agent needs an inbox of its own.
* [/architecture](/architecture) - the organization, pod, and inbox hierarchy, and how a key's scope shapes request paths.
* [/core/send](/core/send) - send email and reply into the same thread.
* [/core/receive](/core/receive) - list what an inbox holds and read messages in full.
* [/core/conversations](/core/conversations) - fetch a whole thread for conversation context.
* [/api-reference](/api-reference/inboxes/list-inboxes) - the full endpoint-by-endpoint HTTP API reference.