Your class stays yours.
Last updated October 15, 2026This policy explains how the Kailah team handles personal information across this website, the waitlist and the research beta. Different actions share different data. A waitlist sign-up does not give us access to your computer or your notes. Questions go to privacy@kailah.ai.
Waitlist eligibility: People aged 13 or older may join the waitlist. Waitlist registration does not provide a Claude account, an AI subscription or guaranteed beta access. Some AI-powered beta builds may have stricter age requirements.
1. Who is responsible and what this covers
The Kailah team is responsible for the personal information it collects through the services described here. AI providers, your school and independent linked websites have their own policies. This policy does not make us responsible for their separate handling. A build-specific research or contribution notice will explain additional collection before it occurs.
We describe both the current waitlist and the beta activities to which you may later agree. The landing page does not implement app telemetry, scanning contribution uploads, billing or an account system. A description of an optional future feature is not a statement that it is already collecting data.
2. Waitlist information and notices
The waitlist is open to people aged 13 or older, including teenagers aged 13–17. Joining does not require a Claude account or an AI subscription, and does not guarantee access to a beta feature that has stricter eligibility rules. When you join, we store your normalized email address, the time it was first added, the waitlist notice version and a “landing” source label. This lets us manage invitations, avoid duplicate entries and record the notice acknowledged. We use the address for beta availability and closely related onboarding updates, not unrelated advertising.
The form uses one Kailah endpoint and our MongoDB database, rather than a waitlist platform. We do not send an email to verify ownership or confirm submission. A successful on-page confirmation means the address was saved or was already listed. Contact us if someone entered your address without permission, you need a correction, or you want to leave.
The form also sends a hidden bot-check field, a form-start timestamp and your acknowledgment of the waitlist terms, privacy notice and provider-access limitation. These are checked during the request and are not stored in the waitlist record, except for the notice version described above. An age-eligibility acknowledgment, if displayed, is for the waitlist and does not verify a third-party AI provider’s age requirements. Joining is separate from accepting beta research conditions, obtaining parental permission where required, and contributing notes.
3. Website requests and abuse prevention
Our hosting and network providers necessarily receive request information such as an IP address, requested path, browser headers and response status to deliver and protect the website. Their operational logs may contain that information. We do not claim that the absence of an analytics script makes browsing anonymous.
For form rate limiting, the endpoint uses a secret-keyed hash of an available trusted network address and a separate hash of the email. IPv6 addresses are grouped by a network prefix. These short-lived counters are separate from the email list. The waitlist endpoint does not store raw IP addresses, user-agent strings or full request bodies in its database or application logs. Hashes are pseudonyms, not guaranteed anonymous information.
Counter records expire through MongoDB’s TTL mechanism, normally after approximately 24–25 hours; removal can lag behind expiry. Without a trusted network-address source, requests conservatively share a network bucket. We also use a service-wide submission cap, validation and bot checks to reduce spam. These measures do not prove an address belongs to its submitter.
4. What is processed on your computer
The companion’s local workflow reads the files you choose, extracts supported text, and builds a local search index. Course records, source text, conversation history and learner context may be stored in the app’s local data directory. The exact storage and supported formats depend on your build.
Local processing does not mean every operation stays on the device. Model requests send selected information to an authorized AI provider, and a separately authorized research submission or note contribution can send information to us. Do not assume that unselected data, backups or archived citation records have been deleted merely because a file is hidden in the interface.
Provider sign-in belongs to the provider. Kailah does not supply Claude accounts or subscriptions, and does not ask you to give our team your Claude password, account cookies or login tokens. Do not include credentials in support messages or diagnostic reports. Joining the Kailah waitlist never authorizes or enables third-party AI access.
5. Data sent to your AI provider
When a permitted provider integration is enabled and you authorize a model request, it may include your question, relevant source excerpts, conversation history and learner context you enabled. Excerpts can contain personal or copyrighted information even if the original file stays on your computer.
The provider receives and handles those requests under its own agreement and your account settings. Retention, safety review and possible model improvement depend on that agreement; local indexing does not opt you out. Review Anthropic’s Privacy Policy and your provider settings before sending sensitive material. Claude consumer accounts are not offered to users under 18. A teenager may join Kailah’s waitlist but must not be directed to create or use an ineligible Claude consumer account. Any beta access for minors must use an independently permitted, age-appropriate configuration, if one is made available. See the Terms on permitted AI-provider access.
6. Browser and companion connections
The browser and companion exchange the information needed for authorized study operations. Some configurations use a local connection; an enabled remote configuration may use transport infrastructure or a relay. Connection and service metadata may be visible to the infrastructure that handles it.
Transport protection depends on the actual build and deployment. We do not make a blanket claim that every route is end-to-end encrypted, that a relay can never access content, or that a local development connection provides production network security. Use only connection methods explicitly supported by your build.
7. Technical and usage data for beta research
The beta helps us understand which devices and workflows work reliably. Before collection is enabled in a participating build, we will explain the research activity and applicable choices. Technical categories can include:
- Computer specifications: operating system and version, processor family or architecture, memory capacity, graphics capability and general device class.
- Software environment: Kailah, companion and browser versions, compatibility state and enabled feature configuration.
- Usage totals: feature events, sessions, imports, page or byte counts, request counts, rough allowance use and provider-reported token totals where available.
- Performance and reliability: timings, throughput, failures, error categories, cancellation and recovery events, and sanitized diagnostics.
We use this information to diagnose bugs, compare compatibility and performance, set practical limits, and improve the product. It may be associated with a beta or device identifier. A combination of device characteristics can identify someone, so we treat it as potentially personal information.
Routine research data excludes source documents, full note contents, full questions and answer text, credentials, precise location, unrelated browsing and unrelated computer files. Local logs are not automatically uploads. If a report or support attachment includes additional content, we need an appropriate disclosure and authorization for that submission. For participants under 18, the applicable beta notice will explain any age-specific protections and when parent or guardian involvement is required. Waitlist sign-up is not research consent, including for teenagers. This landing page does not activate any of those beta collectors.
8. Optional note contributions for scanning
You can choose separately to share selected notes or page samples to improve scanning, OCR, text extraction and document parsing. The contribution flow must show what is being shared and explain any human review, dataset use or model training before upload. For a participant under 18, additional parental or guardian permission may be required and will be requested where applicable. There is no automatic contribution from signing up or merely accepting the general Terms.
If you opt in, the contributed material may include page images or text, layout and extraction output, corrections, annotations, limited technical metadata and the consent record needed to manage it. We use it for the scanning purposes described in that notice. Access is limited to people and service providers who need it for that work. We do not publish identifiable notes or use them to train a general-purpose language model under a routine technical-data acknowledgment.
Remove names, student numbers, grades, personal details and confidential content first. Contribute only material you have permission to share. You can decline, stop future submissions and request deletion of retained source samples. Completed aggregate analyses, software changes and model parameters may not be reversible; any training-specific limits must be explained before that contribution.
9. Purposes and legal grounds
We process information to deliver the services you request, administer invitations, provide support, protect against abuse, improve the beta under the applicable research notice, and meet legal obligations. Where a legal basis is required, we rely on the appropriate basis for that purpose, such as requested service delivery, legitimate interests in security and administration, consent for optional contributions, or compliance with law.
We will obtain consent where required and will not treat a waitlist checkbox as consent to unrelated research, a public dataset or a materially different use. Withdrawing consent does not retroactively invalidate processing that was lawful before withdrawal.
11. Retention and deletion
We keep a waitlist entry while the list remains active and the address is needed for the invitation process, unless you request removal sooner. We review the list when that process ends. Rate-limit counters have the short expiry described above. Hosting logs and backups follow the relevant operational retention and access controls.
For beta research data and optional note samples, the participation or contribution notice will state a retention period or the criteria used to determine it before collection. We do not treat beta participation as permission to retain identifiable material indefinitely. We may retain a limited consent, security or legal record where necessary and lawful, and aggregate information that no longer identifies you.
Uninstalling does not reliably remove local app data, exports, operating-system backups or copies held by a provider. Use the build’s deletion instructions for local records; contact us for records we hold. Deletion requests cannot remove information independently controlled by your provider, but we can direct you to the relevant channel. Backups may take longer to cycle out and are restricted from ordinary use.
12. Access, correction and your choices
Contact privacy@kailah.ai to leave the waitlist, correct an address, ask what we hold, request a copy or deletion, or manage a research contribution. A parent or legal guardian may also contact us about a minor, subject to appropriate verification of authority and applicable rights. Tell us the relevant service and address; do not send passwords or unnecessary identity documents. We may need proportionate verification before disclosing or changing personal information.
Depending on where you live and which laws apply, you may have rights to access, correction, deletion, portability, restriction, objection, withdrawal of consent or appeal, and to complain to a privacy regulator. Authorized agents may contact us where permitted. We will respond within applicable legal periods, explain a refusal where required and avoid penalizing the exercise of protected rights.
California residents may have applicable rights to know, correct and delete information and to opt out of sale or sharing. We do not claim that every statutory business threshold applies to this beta, and we do not reduce rights that do apply. We do not sell or share data for behavioral advertising, including when a Global Privacy Control signal is present.
14. International handling and security
Hosting, database and provider systems may process information in a country different from yours. Where required, appropriate transfer safeguards must apply. Contact us for information about the safeguards relevant to your participation; we do not infer a storage region from the website domain.
We use reasonable technical and organizational controls appropriate to the service, including server-side validation, restricted configuration and rate limiting. No service can promise perfect security, irreversible anonymity or freedom from data loss. Keep backups and avoid submitting material you are not authorized to share.
15. Waitlist age and eligibility for minors
The Kailah waitlist is intended for people aged 13 or older. People aged 13–17 may sign up for waitlist updates, subject to applicable law and parent or guardian permission where required. We do not intentionally collect waitlist information from children under 13 without a legally compliant parental-consent process, which this waitlist does not currently offer. If we learn that an email address was submitted by or for a child under 13, we will take appropriate steps to remove that record and address any applicable obligations. Contact privacy@kailah.ai if you believe this has occurred.
A waitlist entry is not approval for a beta account, an AI-provider account, a research study or note contribution. Individual beta builds may have additional age rules, require appropriate parent or guardian involvement, or be unavailable to minors. AI providers enforce their own age and usage restrictions independently. In particular, current Claude consumer access is not available to people under 18; Kailah does not provide a Claude account or a way to bypass that restriction.
16. Changes and contact
We will update the date when this policy changes and provide appropriate notice of material changes. A new purpose that requires consent will be presented before that processing, rather than being authorized retroactively by an update.
Privacy requests and research choices: privacy@kailah.ai. Technical issues and security reports: support@kailah.ai. General questions for the Kailah team: hello@kailah.ai. Please identify the request in the subject line and include only the information we need to help.