Privacy policy
What is collected, how it supports learning, and who can see it.
What this policy covers
This policy describes the information used by the Chalkchalk pilot, operated by DebatifyAI Inc.. It covers account access, teacher-led classes, personal study, uploaded materials, guided conversations, and the browser coding workspace.
Students and teachers age 13 and up can create accounts. Students can study independently; teachers can prepare coursework and manage their own classes. Student enrollment in teacher-led classes is disabled by default and requires separate activation and agreements. A school may have its own policies or agreements for student records. This page does not establish a school data-processing agreement or certify compliance with education or children’s privacy laws.
Information collected
- Account information: name, email address, account identifier, teacher or student role, and a profile image when provided by your sign-in method. We also store your self-declared 13+ eligibility, any separate 18+ declaration made at recurring-payment review, and the applicable agreement versions and acceptance times. These checkboxes do not verify your actual age, and we do not ask for your full date of birth. You sign in with Google or, where offered, Microsoft or Apple, and your password stays with them; or with an email address and password, which Google Firebase stores and checks for us. Either way, Chalkchalk’s servers never receive your password. For an email and password account we send a link to confirm the address before the account can be used. Apple can provide a private relay email address when you choose Hide My Email. Google Firebase keeps the signed-in account.
- Class and study information: class details, membership, uploaded files, filenames, extracted text and questions, assignment schedules, and processing status.
- Learning activity: messages, attached images, generated guidance, teacher comments, conversation timestamps, saved code, recent saved code revisions, and code or output included in a request for help.
- Voice typing: audio is sent to a speech-to-text provider for transcription. In ordinary chat, the transcription appears in the message box and becomes a saved message if you send it. In essay Voice mode, transcribed speech and relevant draft, selection, planning and recent-command context can be sent to an AI provider automatically to interpret a command; the resulting edits can be saved in your draft and writing history. You can stop listening and undo edits. Chalkchalk does not persist the raw recording in its app database; provider processing and technical logs are subject to the provider terms described below.
- Service information: request and error logs, IP addresses and browser or app details handled by hosting and authentication services, usage counters, processing jobs, and information needed to enforce access and usage limits. The iPhone and iPad app and its Firebase sign-in components also process app interactions and diagnostic information needed for authentication, secure operation and troubleshooting. These operational uses are separate from the optional website analytics choice below.
- Bug reports: the description and optional image you submit, linked to your account identifier, email address and role. Reports also include the page path and screen size so Chalkchalk operators can investigate the problem. Bug reports and their images are stored privately.
- Plan information: if you buy a paid plan, its payment provider, plan, billing period, status, renewal or expiration dates, and purchase consent records. For website purchases we store Stripe’s customer and subscription identifiers. For Apple purchases we store Apple product and transaction identifiers, verified purchase and renewal details, and an account-linked token that connects the purchase to your Chalkchalk account. Payment details stay with the provider; Chalkchalk does not receive your full card number or Apple Account password.
Only include personal information that is needed for the learning task. A photo or uploaded document can contain names and other information about people besides you.
How information is used
Chalkchalk uses this information to sign you in, organize classes, make permitted materials available, extract assignments, provide guided help, save work, deliver teacher feedback, summarize class activity, investigate reported bugs, and operate and secure the service.
Counts and summaries help teachers review the questions students have asked. They should not be treated as a complete picture of a student’s understanding. Summaries can omit images, older messages, or content beyond processing limits and can be inaccurate.
Who can see your work
- Your class teacher can review your conversations, attachments, saved code, and available code history in that teacher’s class.
- Class members can access the course materials shared with their class. Other students cannot read your private conversations through their class membership.
- Personal study is available to the account that owns it and is not automatically shared with a teacher.
- Chalkchalk operators and service providers may process or access information to run the service, provide support, investigate abuse or failures, and meet legal obligations.
Chalkchalk does not publish the textbooks or files people upload, and has no public feed of student conversations. Its Library offers only openly licensed OpenStax textbooks, which a signed-in teacher or student can add to a course; the added copy is kept with that course like any upload. Authorized recipients can still copy or download content they can access, so access controls cannot prevent every form of onward sharing.
Hosting and model providers
Chalkchalk relies on external services. Providing automated help or extracting an assignment can send relevant messages, images, uploaded documents or excerpts, questions, and code context to a model provider. Processing does not all happen on your device.
| Service | Purpose and information |
|---|---|
| Google Firebase and Google Cloud | Authentication, hosted pages, databases, private file storage, server processing, task queues, and operational logs. See Firebase’s privacy information. |
| Resend | Delivery of account emails: confirming your email address and resetting your password. Resend receives your email address, your first name if your account has one, and the one-time link in the email. It does not receive your password or your coursework. See Resend’s privacy policy. |
| Stripe | Payments for website plans. When you buy a website plan, Stripe collects and processes your card or other payment details, email address, and billing address, and handles renewals, cancellations, and invoices. Chalkchalk receives and stores your plan, its status and renewal date, and Stripe’s customer and subscription identifiers, but never your full card number. See Stripe’s privacy policy. |
| Apple App Store | Payments for Apple subscriptions where offered in the iPhone or iPad app. Apple processes your payment details and handles renewals, subscription management, receipts, and refund requests. Chalkchalk verifies Apple’s signed transaction and renewal information and stores the purchase identifiers and subscription status needed to provide the correct allowance and restore purchases. For this verification, we send Apple an account-linked token rather than your Chalkchalk email or raw account identifier. See App Store privacy information. |
| Google Fonts | Delivery of the interface font. Your browser makes requests to Google’s font servers, which receive technical request information. See Google Fonts information. |
| AI model and speech providers | AI processing requires your separate affirmative permission. The current notice in Settings → AI processing identifies the configured recipients and model routes before you allow them. Fireworks AI hosts the model that writes tutor replies, summaries and file readings when that route is configured; Fireworks states that it does not log or store prompts or replies for open models, keeps them in memory only while a request runs (part of a prompt may stay in memory briefly to speed up follow-up requests), and keeps request metadata such as token counts. OpenRouter receives voice recordings for transcription, and any model requests configured to use it, and routes them to model-hosting organizations. Voice recordings go only to Microsoft Azure, which runs Microsoft’s MAI-Transcribe speech model on OpenRouter’s zero-data-retention terms. For other model requests, configured preferred hosts include Wafer, CoreWeave, Together AI, Fireworks AI, DeepInfra, Modal and Makora, with eligible fallback hosts selected by OpenRouter. DeepSeek is a model name, not necessarily the company receiving your data. Direct Anthropic processing and document-reading services appear in the notice when those routes are configured. See Fireworks AI’s data handling information, OpenRouter’s privacy policy, its zero-data-retention routing information, and Anthropic’s privacy information. |
Where the configured route requires zero data retention, that requirement relies on the provider’s contractual and technical commitments; Chalkchalk does not independently audit the provider’s systems. A zero-retention request setting does not establish that every provider log, intermediary system or backup is erased immediately. A provider change or training-enabled configuration requires a separate review and an accurate updated notice. Chalkchalk’s saved learning records remain in its own hosted storage until removed as described below.
Service providers may process information in the United States and other countries where they operate. This pilot does not promise that all data stays in a particular country. Information may also be disclosed when required by law or necessary to address a security incident or rights claim.
AI permission covers the messages, attached images, code, relevant saved conversation and writing context, document pages and voice recordings needed for the AI features you request. It is separate from account terms, upload copyright permission and optional analytics, and is required to enter the learning workspace. You may decline or withdraw and still manage your account, export your data or delete your account. In Settings → AI processing, you can withdraw permission; this stops new AI calls and queued AI processing and cancels active requests where possible. Withdrawal cannot recall information already transmitted. Existing work remains saved until you remove it. A changed provider notice requires a new choice, including for existing accounts.
Storage on your device
Authentication services and the app use browser storage to keep you signed in and remember settings such as theme, onboarding progress, and active file-processing jobs. The standalone coding playground can save code locally on your device. Clearing browser storage can remove those preferences and local work, but does not delete your server-side account or class records.
The iPhone and iPad app also keeps account-scoped notebook and homework writing, imported originals, scanned homework, essay recovery and shared-note copies on the device so interrupted work can be recovered. Signing out preserves those copies for the same account. Account deletion removes owned learning copies from that device; copies on other devices must be removed there or by deleting the app. Purchase-recovery records may be retained separately to reconcile an interrupted Apple payment and prevent duplicate charges. They do not contain your coursework.
Only for signed-in users who have separately declared they are 18 or older, when enabled on the public site and they choose “Allow analytics,” Google Analytics measures visits, demo opens, successful sign-ups, course creation, material uploads, and the start of study sessions using general event labels, cookies and technical information about your browser and device. Google may derive approximate location from network information. We send general page labels rather than class identifiers, URL query strings, or page titles containing your work. We do not send chat messages, code, uploaded content, names, email addresses, or account IDs to Analytics. Advertising features and automatic form, search, and download tracking are disabled. Analytics does not record your screen.
Analytics is optional and is disabled for users under 18 and people who have not separately declared 18+ eligibility. The Google Analytics script is not loaded before an eligible adult opts in. Declaring adulthood at checkout does not opt you into analytics. You can decline or withdraw using the control below or in Settings. Withdrawing stops future collection and removes this integration’s analytics cookies; it does not erase data already received by Google. We remember your choice for this account in this browser for up to six months. The local account identifier used to keep that choice separate is not sent to Analytics, and old browser choices without an account association are ignored. We honor Global Privacy Control and Do Not Track signals. Essential hosting and authentication services still process technical information about requests. Read Google’s privacy policy.
Retention and deletion
Your submitted bug reports and attached images are included in your account data export and removed when you delete your account.
You control your own data. In Settings, Download my data saves a JSON export of your account settings, available learning records, complete saved essay history and owned material metadata. Original files are listed separately and downloaded through authenticated links. Files you do not own, records restricted for a rights claim, provider logs, backups and payment card details are not included. The export identifies withheld records. Delete my account removes your learning access and ordinary account data, including your chats (in your classes and personal study, including classes you left), your uploads and your settings. You can also delete a single chat at any time. A narrowly scoped original source and required case evidence may be kept privately during a documented copyright or legal hold; a deletion receipt identifies that exception and its review date. Necessary payment consent, accounting or legal records may have separate lawful retention periods.
A student’s chats belong to the student. A teacher can read chats in their own class while it runs, but cannot download or delete them. When a teacher ends a class, their coursework, uploaded files and class code are deleted and the teacher loses access; students keep their chats in the ended class, read-only, until they delete them. A teacher’s class download includes only the class details, coursework and roster.
The pilot does not yet delete records automatically at the end of a term. Under a school agreement, the school’s own requests for its records, and deletion at the end of the agreement, are handled with the operator as the agreement sets out.
For access, correction, export, deletion, or a concern about a child’s data, contact jett@chalkchalk.com. Identify the account and request, but do not include your password or send unnecessary copies of sensitive documents. Identity and authority may need to be verified. School-managed records may require coordination with the school.
Applicable law may give you additional rights to restrict or object to processing, receive a copy of your data, or complain to a data protection authority. Service-provider logs, backups, and legal or security records can have separate retention periods; deletion from the app is not a promise of immediate deletion from every provider system.
Students and children
The current pilot accepts student and teacher accounts age 13 and older. Signup requires a declaration that you are at least 13, together with acceptance of these Terms and this Privacy Policy. Users under 18 receive no optional analytics and cannot purchase recurring plans through their own account.
Children under 13 must not create accounts or upload personal information. We do not offer verified parental consent in this pilot. A checkbox, school email, class code or invitation does not supply that consent. If we learn an account belongs to a child under 13, access must be stopped and the information reviewed for removal, subject to necessary identity checks and lawful retention obligations. Report this to jett@chalkchalk.com without sending extra information about the child.
School-managed enrollment requires a separately reviewed authorization and data agreement before it can be enabled. The age declaration does not establish that Chalkchalk is exempt from children’s privacy law; child-directed use requires a separate review. See FTC guidance.
Security and contact
Chalkchalk uses authenticated access, class-level permissions, server-side checks, and usage limits to help protect accounts and stored content. No online system can guarantee absolute security. Keep credentials private and report suspected unauthorized access.
Contact the operator, DebatifyAI Inc., at jett@chalkchalk.com. This page will be updated when practices change; changes that require additional notice or consent will be handled as applicable.