Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.actionlayer.dev/llms.txt

Use this file to discover all available pages before exploring further.

ActionLayer is email infrastructure for AI agents. It gives any AI assistant a real business email identity, a real inbox, conversation threading, a draft submission workflow, human approval notifications, and controlled sending — all through a single REST API or MCP server.

​
What ActionLayer Does

Sender identities

Real addresses on your verified domain. Your AI sends as assistant@yourcompany.com.

Inbox & threads

Inbound mail is parsed, threaded, and stored. Your AI reads conversations in full context.

Drafts & approval

Your AI submits drafts. A human approves before anything sends. Auto-send is opt-in.

Rules engine

Auto-approve trusted contacts, flag urgent words, route by sender — all configurable.

​
What ActionLayer Is Not

  • Not an AI product — we never call OpenAI, Anthropic, or any LLM provider. Your agent is the intelligence.
  • Not email marketing — no campaigns, no bulk send, no tracking pixels. TrackOpens and TrackLinks are always off.
  • Not a Gmail wrapper — you bring your own domain. AWS SES handles inbound; Postmark handles outbound.

​
How It Works

Your AI agent (Claude, GPT, n8n, LangChain, custom — your choice)
         ↓ calls ActionLayer (REST API or MCP)
ActionLayer  ←→  AWS SES (inbound) + Postmark (outbound)
         ↓
   Your verified domain
Your AI is the intelligence layer. ActionLayer is the execution layer.

​
Pick your integration

Quickstart

Sign up, verify a domain, and submit your first draft in five minutes.

MCP server

Native tool use for Claude Code, OpenClaw, Codex, and any MCP-compatible client.

REST API

Full HTTP API for any language or framework.

Features overview

Plain-English tour for non-developers and product-curious readers.