Initiative profile

NoML

WIP

Proposal to add a `noml` directive so content can stay searchable but not be used for machine learning.

Website
noml.info
Latest update
May 28, 2024 Project featured in scholarly article on search and AI opt-out
Primary approach
Preference signal
Pipeline
Collect / Train

What it is

NoML proposes a simple extension to existing robots-style signaling: a noml value that can be added to HTML meta tags or X-Robots-Tag headers. The goal is to let publishers keep content discoverable in search while separately asking that it not be used for machine learning.

It is intentionally lightweight and easy to deploy, but it remains an advocacy proposal rather than a settled web standard.

Watchouts

NoML is a proposal rather than a broadly adopted standard, and it depends on voluntary compliance by crawlers, search engines, and downstream API users.

Evidence trail

Examples

HTML meta tag Source

NoML proposes this as the simplest page-level signal for HTML documents.

<meta name="robots" content="noml">
HTTP header Source

For non-HTML responses, the proposal uses the parallel X-Robots-Tag form.

X-Robots-Tag: noml