IMPORTANT: As of Nov 12 2025, this setting is now default ON for all assignments
The Disable 3rd‑Party Plugins setting blocks text inserts or edits coming from external apps and browser extensions—such as grammar‑enhancing extensions and browser‑based AI tools that modify text.
To keep minor corrections convenient, Rumi allows very small manual pastes or inserts of fewer than 30 characters (≈2–3 words), so routine tweaks don’t become disruptive to normal writing flow.
This setting is especially helpful when you want to:
With the setting ON, attempts by third‑party tools to insert or modify text are blocked, and the student sees a clear message indicating an unauthorized edit occurred. Normal typing and Rumi actions will only continue to work after the student refreshes the page.

With the setting OFF, third‑party edits may be flagged as paste events. However, in some cases these edits may not be detected as "pasted text" depending on how the tool interacts with the browser DOM.


The detection of pasted text depends on how each plugin interacts with the JavaScript page. However, plugins and apps frequently update their logic, what qualifies as “pasted” text can vary based on a plugin’s implementation at any given time. For instance, Grammarly recently updated its algorithm to work across all text inputs, rather than being limited to specific applications like Google Docs.
To solve this, the "Block plugins" feature checks this at a more fundamental level. It evaluates:
a) whether the inserted text is "noticeable" i.e. more than 30 chars (regardless of source e.g. can be Grammarly, Chrome Browser, or Apple AI tools), an
b) whether it was inserted through a Rumi function — such as AI edits, Rumi paste, citation, etc. If it's through Rumi it's an "authorized edit"
If the system detects a noticable unauthorized insertion, it displays a notice to the user and prevents the edit from being sent to the server, so it is never stored.

If you want students to have access to AI for grammar or other enhancements that align with your AI policy we recommend using:
Learn more in the Guided AI and Conversation AI guides.


