Digibrain Gmail MCP

A personal, self-hosted automation tool. Operated by Dāvis Mārtiņš.

This is not a product or service of Acton Academy Riga. Digibrain Gmail MCP is a private tool built and run by Dāvis Mārtiņš for his own use. It is not offered to the public, it is not sold, and it has no connection to the school's operations, staff, families or data. It is published on this subdomain only because the actonriga.org domain is one the operator controls and can therefore verify with Google.

What it is

Digibrain Gmail MCP is a small self-hosted program that runs on the operator's own computer. It implements the Model Context Protocol (MCP), a standard interface that lets an AI assistant call tools. The one thing this tool does is give the operator's own AI assistant access to the operator's own Gmail mailboxes, so the assistant can read and organise mail on his behalf.

What it does

Reads messages, threads, labels and attachments in the operator's own mailboxes.
Organises mail: applying and removing labels, marking read or unread, archiving, moving messages to trash, and drafting or sending replies at the operator's instruction.
Answers the operator's questions about his own inbox, such as summarising what arrived, or finding a message he cannot locate by hand.

Whose mailboxes it touches

Only the operator's own personal and work mailboxes, which he owns and controls. Access is granted by the operator signing in through Google's standard OAuth consent screen and approving the requested scopes for his own accounts.

Who uses it

There are no users other than the operator. There is no sign-up, no account system, no hosted service and no shared installation. Nobody else can connect to it, and there is nothing to connect to: it runs locally, for one person.

Where the data goes

Nowhere new. The tool reads Google's Gmail API from the operator's own machine and hands the results to the operator's AI assistant so it can answer him. OAuth tokens are stored only on the operator's own computer. No mailbox data is sold, shared, published, or used for advertising or for training machine-learning models. The full detail is in the privacy policy.

Contact

Dāvis Mārtiņš — krikauskis@gmail.com