Privacy Policy
Last updated: November 16, 2026
Tutorai is a free tutoring tool for K-12 students, operated by MLN Learning LLC. This policy explains what information Tutorai collects, what it does not collect, and how the information it does collect is used. Tutorai is designed to work without accounts, sign-ups, or the collection of personally identifiable information (PII).
What we don't collect
Tutorai does not require a name, email address, date of birth, school, location, or other personally identifying information to use the tutoring tool. There are no user accounts and no login. The optional contact form and the privacy email address below may receive a message and, if you choose to provide one, an adult reply email. We do not knowingly collect personal information from children, and Tutorai is not designed to identify individual users. Children should not enter personal information in the contact form or in chat.
What stays on your device
Your grade level, subject and topic selections, practice answers, hints used, and mastery progress are stored only in your browser's local storage, lightly obfuscated so the raw values aren't plainly readable. Your full practice history and progress record stay on your device and are not uploaded to Tutorai's servers. (Some of this information is sent in limited form when you use the AI tutor or when anonymous analytics are recorded; see the sections below.) Clearing your browser data, using a different browser, or using a different device will reset this progress, since there is no account to sync it to.
The AI tutor
When you use the AI tutor chat, the message you type and the minimum context needed to answer it (such as the current topic and problem) are sent to our server, hosted by Vercel, which calls Anthropic’s API directly to generate a response. Tutorai does not store chat message text in a database or chat archive; practice progress stays in your browser. Under Anthropic’s commercial terms, API inputs and outputs are generally deleted within 30 days and are not used to train Anthropic’s models without express permission, though content flagged for safety review or required by law may be kept longer (Anthropic organization data retention). Vercel keeps short-lived server logs (which do not contain chat text) under its own terms, typically about one day on our current plan (Vercel Privacy Notice). These providers keep data under their own terms, which can change. Tutorai’s own application error logging records only error metadata (for example HTTP status, error type, request id, and model name), not message text. We ask that you not include personal information in your messages. Messages are filtered and validated before and after being sent. We do not use chat content to identify you, and we do not sell or share it with advertisers.
AI chat is turned off for requests that Vercel geolocates to California (United States). That check uses platform request headers only; Tutorai does not store your IP or region in application data for the block. IP geolocation is imperfect. Chat time limits (about one hour per browser session and two hours per local day) are enforced in the browser using only a date and a minute count in local storage; those counts are not uploaded. Client-side limits can be bypassed. See the child safety policy.
Hosting and service providers
Tutorai is hosted on Vercel, which, like any web host, receives your IP address and basic request information when your browser loads the site or calls our servers. To limit abuse, the AI tutor server keeps a hashed form of your IP address in temporary memory for rate limiting; it is not written to a database. If you email privacy@contact.tutorai.tech or use the contact form, we receive your message and, if you provide one, the parent, guardian, or adult email you enter. Those messages are sent through Resend, our email provider, to a person who reads them. We keep contact-form and privacy emails only as long as needed to respond and for follow-up, and in any event no longer than 90 days, unless a legal duty requires a longer hold. Resend processes the email under its own terms.
Vercel Web Analytics
Tutorai loads Vercel Web Analytics through a same-origin script. Vercel Web Analytics is cookieless. It reports aggregate page-view style metrics and also receives anonymous usage events (event names and non-identifying fields such as topic or intent) so product features like local math answering can be measured in aggregate. To produce these counts, Vercel receives standard request details such as the page address, referrer, browser and device type, and approximate country. Event payloads use the same fixed allow-list as first-party analytics and do not include chat message text. It is not used to identify individual students, and it is not an advertising tracker.
First-party funnel analytics
Tutorai also sends anonymous product events to its own
/api/analytics endpoint. Events are batched in the browser and
posted without a name, account, or persistent user identifier. Each event
includes an event name, a timestamp, and a small data object. The server keeps
only a fixed allow-list of fields and discards everything else. Allowed fields
include subject, topic, grade, whether an answer was correct, learning phase,
mastery, attempts, streak, duration, skill id, grade band, and related
practice metadata; Film Room events may also include tape id, frame, stage,
self-grade, and escalation kind; understanding-check events may include a
check id. Batches are capped (at most 50 events per request). The endpoint
writes the cleaned batch as structured JSON to the hosting provider's
serverless function logs. This data is used only to understand product usage
and improve Tutorai.
Cookies and tracking
Tutorai does not set cookies. It uses your browser's local storage to keep your progress and settings, and a service worker cache so the site loads faster and works offline. Vercel Web Analytics, as noted above, is cookieless. Tutorai does not use advertising trackers and does not sell user data. We do not use Tutorai to build profiles for advertising.
Children's privacy
Tutorai is built for K-12 students and is designed from the ground up to avoid collecting personal information from anyone, including children. Because there are no accounts and no intentional PII collection, Tutorai does not knowingly collect personal information covered by COPPA or similar children's privacy laws. If you believe a child typed personal information into the AI tutor chat, contact us using the address below. Tutorai’s application does not keep a searchable chat log of message text that we can delete on request; any residual copies would be limited to what hosting or the AI provider retain under their systems, which is outside a Tutorai-controlled chat archive.
Data retention
Local progress data is retained only in your browser until you clear it. Application error metadata and anonymous analytics batches may appear in hosting logs for a limited period for debugging and abuse prevention. Tutorai does not operate a separate long-term chat-text store. Contact-form and privacy emails are kept as described under Hosting and service providers.
Changes to this policy
If this policy changes, we'll update the date at the top of this page.
Contact
Questions about this policy can be sent to privacy@contact.tutorai.tech. For other questions, use our contact form.
Read the Terms of Service · Safety protocol · Child safety
Tutorai is operated by MLN Learning LLC.