Privacy policy

Last updated: 2026-08-23

Who we are

Earwise is operated by Earwise LLC, a California limited liability company. Earwise reads text — articles, books, PDFs, and similar — aloud, so you can listen while you walk, drive, or otherwise look away from a screen.

Reach us at hello@earwise.app with any questions about this policy.

What we collect and why

Account information

When you sign up, we store your email address and any profile data your sign-in provider (Google / Apple) hands us. We use this to identify your account across sessions and devices and to contact you about your account if needed.

How we use your email internally: only for account identification, billing, and the emails we explicitly send you (sign-in links, password resets). Our admin tooling that aggregates statistics (active accounts, library usage, pronunciation flags) does not query or display individual email addresses — we work from anonymized user identifiers (UUIDs) so the operator running the dashboard never sees, for example, "person at example.com listened to Pride and Prejudice today." If you want zero-email signup, use Apple Sign-In with "Hide My Email" — we get a relay address and never see your real one.

Content you save

Earwise is a "save articles for later" service. When you save an article, book chapter, or PDF, we store on our servers:

• The URL or source you saved.
• The article title and author when available.
• The extracted text content — we have to store this so we can read it back to you.
• Images and figures contained in documents you upload. These are stored privately under your account, so a document you upload on one device is complete on your others. They are never shown to other users and never placed in the shared audio cache described below. To read a figure's description aloud, an image may be sent to our AI provider (see "Who we share data with" below) for a one-time, one-sentence description; the description text is then stored the same way the rest of your document's text is.
• Per-sentence audio files generated by our text-to-speech provider. Each audio file is stored in a shared cache keyed only by a cryptographic hash of the exact sentence text — never under your account, your email, or any user identifier. Hashing is one-way: we cannot reverse the hash back into the original sentence, and we have no record of which sentences belong to which users. The only path through which a different user receives the same audio file is if they independently submit a sentence with identical characters (punctuation and spacing both matter), at which point our system recognizes the existing hash and serves the cached audio instead of re-synthesizing. Your saved articles, books, and PDFs remain private to your account.

When you delete an item from your saved list, the associated content is removed within 30 days. Generated audio that's content-addressed and shared across users is not deleted individually, but the audio bytes are not linkable to your account after your save row is removed.

Usage data

We track your monthly synthesis volume (characters of text we've converted to audio for you) so we can enforce per-account character caps. We do not maintain a per-page reading-progress profile beyond what's needed to resume playback where you left off.

What we DON'T collect

• We don't track what you browse. The browser extension only reads pages you actively click to save — never pages you merely visit.
• We don't run third-party analytics, advertising trackers, or behavioral profiling tools.
• We don't share your saved content with other users. Your articles, books, and PDFs stay in your account, visible only to you. The audio files we generate are stored in a shared cache keyed only by a hash of the sentence text — with no link back to which user originally requested them. If another user later submits identical text, they receive the cached audio for free, but they have no way to know it came from your content originally.

Browser extension specifics

Earwise ships browser extensions for Chrome (desktop) and Safari (macOS and iOS). Both follow the same architecture and privacy model. The permissions below apply to both:

• activeTab: read the URL and title of the tab you're on at the moment you click the toolbar button, context menu, or keyboard shortcut. This permission does NOT allow background tab monitoring — only the user-gestured tab is exposed.
• scripting: run a single, on-demand extraction script in the tab you actively click to save. The script reads the article you're already viewing, extracts the text, and sends it to Earwise's server so it can be saved to your queue. It runs only on user gesture (toolbar click, context menu, keyboard shortcut) — never in the background, never on tabs you don't actively save.
• storage: store your sign-in token locally, sandboxed to the Earwise extension.
• contextMenus: register "Save to Earwise" right-click items.
• host_permissions: https://earwise.app/*: the only host the extension is permitted to make network requests to. The extension cannot send your data anywhere else.

Paywalled content and subscriptions

If you're a paid subscriber to a publication (NYT, WSJ, FT, Bloomberg, Substack, etc.), your browser is already authenticated and has fetched the article on your behalf when you click "Save to Earwise." The extension reads the page your browser already rendered — Earwise's server never sees your subscription credentials, cookies, or login session. We never bypass paywalls, never store third-party login credentials, and never share saved content with other Earwise users. Your saves are yours.

If you're saving from a public page, the extension still uses the same pattern — we read what your browser rendered locally. For certain content types (browser-rendered PDFs), the extension falls back to having our server fetch the URL directly, but this fallback never includes your browser cookies.

Who else processes your data

We rely on these third-party services. Each is bound by a data processing agreement that limits use to providing the service to us:

• Supabase (database, authentication, file storage)
• Vercel (web hosting and serverless API)
• Google Cloud Text-to-Speech (text → audio synthesis)
• Anthropic (AI text cleanup and document summaries; summaries may be generated automatically in the background for documents you save, in addition to when you open one directly. This also covers one-sentence descriptions of images and tables in your documents, generated once per figure so they can be read aloud — an image sent for this purpose is used only to generate that description and is not retained by us for any other purpose beyond what is described here)
• Stripe (if you become a paying subscriber)

Your rights

You can access, export, correct, or delete the data we hold about you at any time. For accounts in the EU, UK, California, and other jurisdictions with applicable laws, you have a legal right to do this, but we extend it to all users by default. Sign in to your account and use the in-app settings, or email hello@earwise.app and we'll handle it within 30 days.

Children

Earwise is not directed at children under 13 and we do not knowingly collect data from anyone under 13. If you believe a child has provided us data, contact us and we will delete it.

Changes to this policy

We update this page when our practices change. Material changes get a banner in the app and an email to active users. Substantive changes are dated above.