LearnLayer Privacy Policy

Last updated: July 8, 2026

The short version: nothing leaves your browser until you click a lesson button. When you do, the AI answer you are looking at plus your question for it are sent once, used once to generate that lesson, and not stored. No accounts, no ads, no data sales, no tracking of what sites you visit.

What LearnLayer reads, and where

The extension's content script runs only on the AI chat sites listed in its manifest. On those pages it reads the page text locally, inside your browser, purely to figure out where an answer ends so it can place its buttons. That local reading never leaves your machine. LearnLayer has no permission to see any other website you visit; the browser enforces this, it is not a promise we could quietly break.

What is sent, and when

Exactly one kind of network request exists: when you click Why?, How?, Teach Me, Quiz Me, or Make Me Do One (or press the button in the web app or CLI), LearnLayer sends:

the AI answer you are asking about (capped at 8,000 characters), and
your question that produced it, when available (capped at 4,000 characters), and
which of the five modes you picked, and
a random device id (see "The free-tier counter" below) used only to count your daily free lessons.

That is the entire payload. No account id, no cookies, no browsing history, no page URL.

What happens to it on our server

Our server passes that text to Anthropic's API to generate the lesson, returns the lesson to you, and is done. The text is not written to any database, key-value store, or content log. The only thing persisted is aggregate metering (a count of tokens used and their cost, tagged "learnlayer"), which contains none of your text.

The one honest caveat

Anthropic's API is a real third party in this pipeline: it must see the question and answer text to generate the lesson, the same way any AI product works. We will never claim "nobody ever sees your data" in an absolute sense. The accurate promise, and the one we make, is: we see it once, use it once, and do not keep it. Anthropic's handling of API data is governed by their privacy policy; API inputs are not used to train their models per their commercial terms.

The free-tier counter

To enforce the 10-lessons-a-day free tier, LearnLayer generates a random device id on first use (a string like ll_ followed by random characters) and stores it on your device (chrome.storage.local in the extension, localStorage in the web app, ~/.learnlayer-device for the CLI). It is sent with each lesson request so our server can keep a per-device daily count. That count is the ONLY thing stored against the id: a number that expires within two days. The id is pure randomness: it is not derived from your hardware, email, or anything about you, it is not linked to any account, and it is never used for tracking or shared with anyone. Deleting it simply makes your device look brand new. If a request arrives without a device id, the daily count is kept against the request's IP address instead (also expiring within two days).

What we never do

No selling or sharing of data with anyone, for any reason.
No advertising, no ad networks, no third-party trackers or analytics scripts in the extension.
No account or sign-up required for the free tier.
No reading of pages outside the AI chat sites in the manifest.
No storing of your questions, answers, or lessons server-side.

Rate limiting

Our server applies a short-lived, per-IP rate limit to keep the free service healthy. IP addresses are used transiently for that counter (which expires within minutes) and are not tied to your content or kept as logs of who asked what.

Changes

If this policy changes in a way that affects what is collected or sent, the change will be listed here with a new date, and the extension will not silently expand what it sends.

Contact

Questions: support@draftguardapp.com