Skip to main content

What is Openclaw?

Openclaw is an open-source AI agent framework that runs on your machine 24/7. It connects to messaging platforms (WhatsApp, Telegram, Slack, Discord) and executes tasks autonomously using Tools and Skills. By giving your Openclaw agent a Commune email inbox, it can send outbound emails, receive inbound emails via webhook, and manage threaded conversations — all programmatically.

How it works

  1. Create a Commune inbox for your agent via API
  2. Set up a webhook endpoint so Openclaw receives inbound emails
  3. Create a SKILL.md that teaches Openclaw how to use the Commune API
  4. Your agent can now send, receive, and reply to emails autonomously

Setup guide

1

Create an inbox

This gives your agent an address like my-openclaw-agent@agents.commune.email.
2

Configure a webhook for inbound emails

Set up a webhook on your inbox so Commune forwards inbound emails to your Openclaw instance.
If running Openclaw locally, use a tunnel like ngrok or cloudflared to expose your webhook endpoint.
3

Add the Openclaw webhook hook

In your openclaw.json, add a hook that listens for the email webhook topic:
4

Create the Commune email SKILL.md

Skills are markdown playbooks that teach Openclaw how to perform tasks. Create this file in your Openclaw workspace:
5

Set environment variables

6

Test the integration

Send a test message to your Openclaw agent via any channel (WhatsApp, Telegram, etc.):

Advanced: Auto-reply to inbound emails

You can configure your Openclaw agent to automatically process and reply to inbound emails. Add this to your SKILL.md:

Resources

Openclaw GitHub

Openclaw source code and documentation

Openclaw Hooks

Webhook hook configuration docs

Messages API

Full Commune messages documentation

Webhooks

Webhook setup and payload format
Last modified on March 19, 2026