Privacy Policy

StreamTutor (“we”, “us”) is a Chrome extension and website that helps you learn languages from your own Netflix. This policy explains what we collect and why. We aim to collect as little as possible.

What the extension does on Netflix

The extension runs in your browser, only on Netflix. It reads the subtitle files your Netflix player already loads and displays dual subtitles. It does not change which audio track plays and does not modify Netflix's service.

When you click a word, the extension sends that word, its subtitle line, the two lines before and after it, and your two languages to our backend. Our backend uses OpenAI to write the explanation and Google (Gemini text-to-speech, through Vertex AI) to speak it; if Google's service is unavailable, OpenAI speaks it instead. Google receives only the text to be spoken and the names of your two languages, never your account, email or IP address. To show what words mean and how they sound, the extension also sends single words and lines from the subtitles.

Microphone (Say it). StreamTutor uses your microphone for one thing: Say it, where you say a subtitle line out loud. It starts when you hold Space on a line or click Say it, and stops when you let go or stop talking. While it's on, your voice is streamed through our backend to OpenAI, which transcribes it live, and once more when you finish, so the extension can light up the words you said. Only the line's language goes with it, plus a pseudonymous ID (never your name or email) that helps OpenAI detect abuse. We don't store recordings or transcripts. OpenAI doesn't use your voice to train its models and may keep it for up to 30 days to monitor abuse. The microphone may stay ready for a few seconds after a try so the next one starts at once; nothing is sent then. Before version 1.1, StreamTutor also used the microphone while you held Space to ask the teacher a question; that recording went to our backend and OpenAI and wasn't kept. Voice questions end with version 1.1.

What we store

If you create an account, we store your email. Your password is handled by our sign-in provider, Supabase, and the extension keeps only a session token. We also store your learning progress, so you can review it and pick up on any device:

• the words you learn, each with the lines you met it in (the show and the moment, its meaning there and the teacher's explanation), and your review answers and schedule;
• the Netflix episodes you watch with StreamTutor on, the most recent 60: the title, the episode, how many minutes you watched and its words, so your progress in each episode follows you;
• the lines you had explained;
• if you asked the teacher questions before version 1.1, a short label for each and her answer. Version 1.1 stops saving them, and we'll delete them all 30 days after it's released;
• how many word explanations you've used today and how many Say it keys we've issued you, for your daily allowance and to keep costs in check.

These records are tied to your account: no other user can see or change them. Without an account nothing is synced, and your words stay in your browser.

If you enter your email on a phone to get the link for your computer, we use it only to send that link and one reminder, and delete it after 30 days.

We cache generated explanations, audio and word meanings, keyed by subtitle text and language pair (and, for a word, the word itself), so popular content doesn't need to be regenerated for every user. The cache is shared across users and holds no personal data: the subtitle line, the word and what the teacher says about it, never who it was for. The audio of a single word said on its own, played on review cards, is generated once per language and word and cached the same way. Your sign-in token is only used to check your plan and count your daily allowance.

When the extension can't read a title's subtitles, it reports that to our server logs: the Netflix title ID, whether its subtitle reader had started, and the extension version. It carries no account or user ID, and is sent at most three times a day. Our servers see your IP address, as any server does; we use it in memory to limit request rates and don't store it.

From version 1.1, the extension also sends us product events, so we can see which parts of StreamTutor help people learn: for example that you opened a Netflix title, clicked a word, started a review or tried Say it, and how many words of the line you said. Each event holds its time, a random ID for your install, your account ID when you're signed in, and a few small details: the language you're learning and your own, your browser's language and its time zone offset from UTC, your plan, the Netflix title ID, which part of the extension it came from, counts, durations and the extension version. Never subtitle text, the words themselves, your email or your voice. A daily job deletes product events once they're 180 days old. You can stop sending them in the extension's settings: turn off “Share usage data to help improve StreamTutor”.

What we don't do

We don't sell your data or use it for advertising. The only viewing history we keep is the one above: the last 60 Netflix episodes you watched with StreamTutor on, with their title and minutes watched, in your own account, and the Netflix title IDs in product events. We know nothing about what you watch without it. We don't store subtitle content beyond what's needed to cache explanations and to keep your words with the lines you met them in, and we don't keep recordings of your voice or transcripts of what you say.

Third parties

We use OpenAI to write explanations and to transcribe what you say in Say it, Google Cloud (Gemini text-to-speech, through Vertex AI) to turn explanations and single words into speech, with OpenAI as a backup, Supabase for accounts and data storage, Stripe for payments, and Vercel for hosting. You pay on Stripe's own checkout page, so we never see your card details. These providers process data on our behalf under their own terms. OpenAI doesn't use data we send through its API to train its models, and may keep it for up to 30 days to monitor abuse. Google doesn't use data we send through Vertex AI to train its models.

Your choices

You can use the extension's dual subtitles without an account. Say it is optional: StreamTutor needs your permission before it uses your microphone, and you can turn it off in the extension or in Chrome's settings. You can delete saved lessons at any time, and you can request deletion of your account and associated data by emailing us. Deleting your account also deletes your words, review schedule, episode progress and the product events linked to your account (and any questions you asked the teacher before version 1.1), and we'll delete any of them on their own on request.

Your rights

We use your data to give you StreamTutor (your account, your words and progress, explanations and Say it), which is our agreement with you, and for our legitimate interest in keeping the service working, affordable and free of abuse (daily allowances, rate limits and logs) and in improving it (product events, which you can turn off). We keep your account data for as long as you have an account. When you request its deletion, we delete it within 30 days. Our providers may process data outside the European Economic Area, including in the United States, under their own data protection terms.

You can have us show you, correct, export or delete your data, and you can object to how we use it. You also have the right to complain to a data protection authority: in the Netherlands, that's the Autoriteit Persoonsgegevens.

Children

StreamTutor isn't meant for children under 16. If you're under 16, check with a parent or guardian first.

Who we are and contact

The data controller is Streamtutor, Vechtstraat 101, 1079 JC Amsterdam, the Netherlands. Anything about your privacy, or a request to see or delete your data? Email hello@streamtutor.ai.