Introduction
Montara Mail is an AI mail client that runs your own model, on our servers or on yours.
What it is
Montara Mail puts every mailbox you use in one place: Gmail, Outlook, Yahoo, iCloud, Lark, Zoho or any IMAP server. A background sync keeps each one current, and an AI agent sorts new inbox mail, summarizes long threads and drafts replies for you to review.
You stay the one who sends. The agent prepares replies, forwards and new mail; nothing leaves your mailbox until you approve it.
Your model, your server
Each organization picks the model its AI features use: any OpenAI- or Anthropic-compatible endpoint, or a model you host yourself, such as Ollama. The whole product is open source under GPL-3.0 and ships as one Docker image, so you can also run the server yourself.
Your email data is used only to provide the service and is never sold.
Where to start
Getting started
Sign in, create an organization and connect your first mailbox.
Connect mailboxes
App passwords, OAuth, and what the sync keeps in step.
Self-hosting
Run the API and web app on your own server with Docker.
Connect MCP clients
Use your mailboxes from Claude Desktop, Cursor and other MCP clients.
How it fits together
| Part | What it is |
|---|---|
| Web app | The mail client at app.montaramail.web.id, a single-page app. |
| API | A Go server: sign-in, organizations, IMAP sync, SMTP send, AI features and the activity log. In production it also serves the web app. |
| Your model | The LLM provider each organization configures, called by the API. |
| Your mailboxes | Read over IMAP and sent through over SMTP, with your app password or an OAuth sign-in. |