This document describes how the code in this repository actually behaves, which makes it an accurate starting point rather than filler. It has not been written or reviewed by a lawyer, and it is missing jurisdiction-specific obligations you are likely to have — GDPR, UK GDPR, CCPA/CPRA, and COPPA in particular, the last of which matters a great deal for a product aimed at students who may be under 13.
Every TODO below must be filled in, and the whole thing reviewed by a qualified lawyer, before you take a single real payment. Chrome Web Store review also requires a published, reachable privacy policy that matches what your extension actually does.
Privacy Policy
Last updated: TODO — insert publication date
1. Who we are
Answr ("we", "us") provides a browser extension and related website that help students read and work through questions displayed on their own screens.
Legal entity: TODO — registered company name
Registered address: TODO — full postal address
Data controller contact: TODO — privacy contact address
2. The short version
Screenshots never leave your device, on any plan. Text recognition runs inside your own browser using a WebAssembly OCR engine bundled with the extension. There is no image upload endpoint in this product, so there is nothing for us to receive, store, or lose.
The recognised text is transmitted. Once the extension has extracted text from the screenshot on your machine, that text is sent to our server and forwarded to our AI provider so it can be answered. This applies to every scan, on every plan. The image is not sent — only the characters read from it.
3. What the extension processes locally
- Screen captures. When you press the hotkey, the extension captures the visible area of the current tab. The image is held in memory in your browser, cropped if you select a region, passed to the local OCR engine, and discarded. It is never transmitted.
- Recognised text. Extracted in your browser and displayed in the popup. Not transmitted by the extension as shipped.
3a. Question text sent for AI answering (Premium and Lifetime)
For every scan, the extension sends the recognised text to our
/solve endpoint, which forwards it to OpenAI
(privacy policy,
API data usage)
to generate an answer. What this means in practice:
- The screenshot is not sent. Only the text extracted from it.
- Your licence key accompanies the request so we can confirm your plan.
- This applies to every scan, on every plan — Pro, Premium and Lifetime alike.
We reach OpenAI through their API, not the consumer ChatGPT product. OpenAI does not use data submitted through the API to train its models by default. This is a different arrangement from ChatGPT, where conversations may be used for training unless you opt out.
TODO — verify before publishing. Confirm the current wording of OpenAI's API data-usage terms and their retention period for API requests, then state both here in your own words. Do not publish this paragraph on the strength of a summary; read the terms.
TODO — state whether you log question text, and for how long. The shipped code does not persist it, but your hosting provider's request logs may capture it. Decide, then say so plainly here.
Summarising a page
When you press Summarise this whole page, the extension reads the text of the page you are looking at and sends it to our server, which passes it to the same AI provider described above. This is more information than a scan sends: a scan transmits only the words inside the box you selected, whereas a summary transmits the readable text of the entire page, including parts you have not scrolled to.
It happens only when you press that button, and only for the tab you are looking at at that moment. The extension does not read pages in the background, does not read other tabs, and does not read pages you have not asked it to summarise. Before sending, it removes scripts, styles, navigation, headers, footers, sidebars and forms, and sends at most the first 12,000 characters of what remains. Images and screenshots are never sent on this path either.
Do not summarise pages containing information you would not want transmitted — webmail, online banking, medical portals, or anything behind a login that holds personal records about you or anyone else.
TODO — if you later change solveQuestion() in
extension/solver.js to route more categories to a provider, or change
providers, this section must be rewritten before that version ships.
4. What we receive on our servers
- Your license key, sent to our
/verifyendpoint when you open the extension popup and once daily in the background, so we can confirm your subscription is active. - A device identifier — a random value the extension generates the first time it runs and stores on your own device. It is sent with your license key so we can enforce our one-device-per-license rule. It is not derived from your hardware, your browser, or anything about you: it is a random number that means nothing outside your license, cannot identify you, and is not used for tracking, analytics, or advertising. Signing out of the extension erases it from our records.
- Standard server request data — IP address, timestamp, and user agent — logged when your browser contacts us. We use IP addresses for basic rate limiting.
- Purchase records — the tier you bought, your Stripe customer and subscription identifiers, the email address you gave at checkout, and the license key we issued.
We do not receive or store your card details at any point. TODO — state your log retention period.
5. Payment processing
Payments are handled by Stripe, Inc., which acts as an independent controller of the payment data you give it. Their handling is governed by Stripe's own privacy policy at stripe.com/privacy.
6. Legal basis and purposes
TODO — if you have users in the UK/EU you must state a lawful basis for each purpose (contract performance for licensing; legitimate interests for fraud prevention and rate limiting; consent for any analytics or marketing email).
7. Sharing
We do not sell personal data. We share it only with processors needed to run the service:
- Stripe — payments and subscription management.
- OpenAI — generates answers for non-maths questions. Receives the recognised question text; receives no image and no payment data. API submissions are not used for model training — see section 3a.
- TODO — hosting provider
- TODO — email provider, if you send transactional mail
- TODO — analytics provider, if any; if you add one, this policy needs a cookie/consent section it currently does not have
8. Retention
License records are kept for as long as the license exists and for TODO — period afterwards for accounting and tax purposes. Server logs are kept for TODO — period.
9. Your rights
Depending on where you live you may have rights to access, correct, delete, port, or restrict processing of your personal data, and to object to it. To exercise them, contact us at TODO — privacy contact address. We respond within TODO — statutory period for your jurisdiction.
10. Children
TODO — this section needs real attention. A study tool will attract users under 13, and both COPPA (US) and the UK Age Appropriate Design Code impose specific obligations on services likely to be accessed by children. Decide and state your minimum age, how you enforce it, and whether you require parental consent.
11. International transfers
TODO — name the countries data is processed in and the transfer mechanism (e.g. SCCs) if you serve UK/EU users.
12. Changes
We will post any changes on this page and update the date above. Material changes will be communicated TODO — describe how.
13. Contact
Questions about this policy: TODO — privacy contact address. For product help, see the support page.