Skip to main content

Getting Started

mailbot gives you a fast way to create a programmable inbox and send your first email without learning internal infrastructure details.

For beta, the recommended first-success path is:

  1. sign up
  2. create a sandbox inbox
  3. send one real email via API or SDK

Start with the API / SDK flow first. MCP is available as a follow-up integration path, not the primary beta onboarding path.

Current public/demo surfaces:

  • Docs: https://getmail.bot/docs
  • Dashboard: https://getmail.bot/dashboard
  • Console: https://getmail.bot/dashboard/console

Node.js​

import { MailBot } from '@yopiesuryadi/mailbot-sdk';

const client = new MailBot({ apiKey: 'mb_xxx' });
const inbox = await client.inboxes.create({ username: 'support-bot' });

console.log(inbox.id);
console.log(inbox.address);

If you later use raw API examples, replace INBOX_ID with the real inbox.id returned here.

Python​

Python package is published as mailbot-sdk.

from mailbot import MailBot

client = MailBot(api_key="mb_xxx")
inbox = client.inboxes.create(username="support-bot")

The sandbox path is enough for first success. You do not need custom-domain DNS setup before sending your first real message.

Agent Frameworks​

If you want to hand inbox actions directly to an AI agent, the fastest path today is the LangChain package:

pip install mailbot-langchain langchain langchain-openai
from mailbot_langchain import MailBotToolkit

toolkit = MailBotToolkit(api_key="mb_xxx")
tools = toolkit.get_tools()

The toolkit defaults to https://getmail.bot/v1 and includes inbox, message, thread, webhook, domain, compliance, and engagement tools. For deeper examples, see examples/langchain-ai-sdr-agent.py, examples/langchain-ai-support-agent.py, and examples/langchain-ai-onboarding-agent.py in the repo.