---
title: "Rules"
description: "What happens to mail as it arrives."
url: "https://openemail.uk/docs/mcp/tools/rules"
area: "MCP server"
category: "Tools"
---

# Rules

What happens to mail as it arrives.

## Rules tools

| Tool | What it does |
| --- | --- |
| listRules | Every rule on the active connection, in the order they are evaluated, with what each one does and how often it has fired. |
| getRule | One rule in full: its conditions, its actions and whether it is on. |
| createRule | Write a rule: conditions over an arriving message, and what should happen to anything matching them. The rule is created DISABLED. |
| testRule | Which of the last thirty days’ messages a rule would have caught, and what it would have done to them. Changes nothing. |
| setRuleEnabled | Turn one rule on or off, once a person has read it back. |

> `createRule` writes the rule DISABLED, which is the one place this server deliberately differs from the REST API. `POST /rules` defaults to enabled. A rule that archives, trashes or forwards mail should not start doing it on the strength of a chat message: write it, call `testRule` to show the person you are working for which of their own messages it would have caught, and only then ask for `setRuleEnabled`.

> There is no updateRule, no deleteRule and no reorder, for the same reason there is no template deletion: those are the operations that break somebody else’s setup from inside a chat window. A deleted rule cannot be restored, and a reorder silently changes what every future message does. All three are in Settings → Rules and on the REST API, where a person is the one clicking.

> Rules belong to the CONNECTION, so these act on whichever mailbox `setActiveConnection` last named, and a rule a colleague wrote is one of them. A mailbox holds 100 rules, each with at most 20 conditions and 10 actions.

> Nothing here is retroactive. A rule decides what happens to mail arriving after it is enabled; it does not sweep a mailbox that already exists, and `testRule` reports rather than files.
