Privacy Policy

The app collects information you provide directly: children's names, ages, characteristics, and descriptions of situations, your rating of each moment (it got worse, about the same, it helped, or it helped a lot), and an optional note. This information is stored on your device. If you separately choose to send anonymous insights, we receive only the summary described under “Weekly digest and anonymous insights”.

When you use the microphone, your voice is sent to OpenAI's gpt-transcribe model and transcribed to text transiently so the app can generate guidance. We do not retain your recordings. Nothing you say is used to train models.

Information stored on your device is used to give you personalized parenting guidance and, if you turn it on, a digest calculated on the device. A guidance request is sent to our functions on Netlify. That request includes only the child you asked about: the story text, the context you chose (location, your mood, who else was there, and how urgent it feels), that child's profile (name, birth date, gender, intake answers, notes, and whether a new baby recently joined the family, without the baby's name), and that child's eight most recent moments (each moment's story, a short summary of the earlier guidance and its advice headlines, your rating, and, when saved, when and where it happened and your mood). It also includes a RevenueCat ID so the function can check the subscription. Names, profiles, and history of other children are not sent. If an older version of the app still includes them, the function leaves them out of the prompt. The optional moment note stays on your device and is not included in the request. OpenAI's gpt-6-luna model then receives the story, that context, that child's profile, and up to five of those recent moments. The RevenueCat ID is not part of the text sent to the model. We do not store those prompts.

We do not sell personal information. We do not show ads in the app or on this site. We advertise Kidsit on Meta's platforms and, with your consent, measure those ads using the Meta Pixel. The providers below match what the app and this site actually use:

• OpenAI: gpt-6-luna writes guidance and gpt-transcribe transcribes voice, as described above. OpenAI's privacy policy applies to that processing.
• Netlify: hosts the guidance and transcription functions. Those requests pass through Netlify. Netlify's logs can include an IP address. Anonymous insights are not sent to Netlify, and we do not copy that IP into the insights database.
• RevenueCat: checks whether a subscription is active. The app sends RevenueCat an app user id, and the Netlify function can send that same id when it checks the entitlement. It also holds the Android install source and the free-answers counter described under “App usage counts, install source and free answers”. Apple or Google bills you. We never see your card number.
• Insert Affiliate: if you open the app from an affiliate link, an affiliate identifier is stored on the RevenueCat subscriber. It is not sent with anonymous insights.
• Vercel: hosts this website (kidsit.ai) and the endpoints for anonymous insights and funnel counts. Vercel request logs include IP addresses for the retention period of the plan. We do not copy the IP address or the request body into our database.
• Neon: Postgres, connected to our Vercel project, stores anonymous insight submissions you choose to send and the daily funnel counters. See “Weekly digest and anonymous insights” and “App usage counts, install source and free answers”.
• Google: Google Analytics 4 and store-click measurement on this website only, after you accept cookies. They do not receive what you enter in the app. Details are under “Cookies and local storage”.
• Meta Platforms: the Meta Pixel on this website only, after you accept analytics cookies, to measure our ads and site visits. It does not receive what you enter in the app. Details are under “Cookies and local storage”.

Moments, ratings, notes, and child profiles are stored on your device. We do not store those moments or profiles on our servers. If you delete the app, that on-device information is removed with it. The exception is anonymous insights you choose to send: the counts and the optional rating are kept, and free text is deleted after 90 days. Funnel counts are kept only as daily totals.

The daily or weekly digest is built on your device from moments already stored there, separately for each child. It is off unless you turn it on. Building or reading the digest does not send those moments to us.

Sharing anonymous insights is a separate choice, and it is also off by default. Nothing is sent unless you turn sharing on and then confirm one send, after a preview of exactly what will leave the device. Turning either option off does not change the price or the guidance you can use.

If you send a summary, it can include counts of how moments went (helped a lot, helped, about the same, got worse, or not rated) grouped by a fixed list of situation types; counts of which guidance techniques appeared in moments that helped or did not; an optional rating from 1 to 5 of how much Kidsit helped that week; optional free text you write, up to 500 characters; the app language; whether the device is iOS or Android; and the app version at major.minor only. A daily digest is still stored as an ISO week (for example 2026-W44), not the exact day. In the digest wording on the device, “about the same” is counted with moments that did not help. The anonymous counts keep the original rating, including “about the same”, and do not fold it into another number before it is stored.

We do not receive children's names, stories, AI replies, notes, ages, birth dates, gender, diagnoses, how many children you have, or any user, device, install, or RevenueCat identifier. The row is stored under a random id. The data is not linked to you, and your IP address is not stored in this database.

Before sending, the app removes names of children and siblings saved on the device, email addresses, phone numbers, web addresses, 9-digit ID numbers, and full dates from the free text, and shows you the cleaned version. The server removes emails, phone numbers, web addresses, 9-digit ID numbers, and full dates again before writing. Do not write names, places, contact details, or medical details. The text is stored as it remains after that cleaning.

Free text is deleted automatically after 90 days. The counts may be kept longer. Because the send is anonymous, we cannot find or delete one specific submission. To remove a piece of free text that is still inside those 90 days, email us and quote it.

Submissions are stored in Neon Postgres through Vercel, and the endpoint runs on Vercel. Vercel request logs include IP addresses for the plan's retention period. We do not write the request body or the IP address into our database, and we do not use those logs to identify you.

The purpose is to improve Kidsit. You can turn sharing off at any time in the app's settings, and the send button then disappears.

Anonymous funnel counts. The app reports when an install reaches certain steps for the first time: finishing onboarding, adding a first child, sending a first moment, receiving a first answer, rating a moment, seeing or closing the subscription screen, and buying a subscription. Each step is sent once per install. A report contains only the step name, whether the device is iOS or Android, the app language, the app version at major.minor only, and, on Android, the install source (the utm_source value from the Google Play link, if there was one). It contains no user, device, install, or RevenueCat identifier, no child data, no story text, and no cookies. The server adds the day (UTC) and keeps only a daily counter for each combination; it does not keep a row per report. The counters are stored in Neon Postgres through Vercel. Vercel request logs include IP addresses for the plan's retention period, and we do not write the IP address into the database. There is no setting in the app to turn these counts off, and they are not linked to you.

Android install source. If you install the Android app from a Google Play link that carries campaign parameters (for example, from this website), Google Play gives the app those parameters once after installation. The app keeps only the utm_source, utm_medium, utm_campaign, and utm_content values and drops any ad click IDs. It stores them on your device and on your RevenueCat subscriber (as media source, ad group, campaign, and creative), linked to the RevenueCat app user id, so we can see which channels lead to subscriptions. The App Store has no equivalent, so the iOS app does not do this.

Free answers counter. Your first 3 guidance answers are free. To count them, our Netlify function stores one number, how many free answers have been used, as an attribute on your RevenueCat subscriber, linked to the RevenueCat app user id. It does not include the story, the answer, or anything about your child. Subscribers are not counted.

We use essential first-party storage so the site works (your language choice and accessibility-menu preferences). We also keep how you arrived on this visit (the source, such as a search engine, an AI chat or an ad, its campaign parameters, any ad click ID, and the landing page) in session storage only, which is cleared when you close the tab, and add the source and campaign to the App Store and Google Play links, to measure which channels lead to app downloads. With your permission we also load Google Analytics 4 to understand how the site is used. After you accept, Google's tags may also send a conversion measurement ping when you click an App Store or Google Play button (used to attribute store visits from this site). Analytics and those pings do not receive anything you enter in the app. Google tags do not load, and non-essential cookies are not set, until you accept.

Meta Pixel. With the same permission we also load the Meta Pixel from Meta Platforms (Facebook) to measure how well our ads work and to count site visits. It loads only after you allow analytics cookies in the banner, and you can withdraw that any time under “Cookie settings”. When it runs, Meta may receive cookies such as _fbp and _fbc, browser and device information, the pages you visit on this site, and your IP address. It may also send automatic click events when you click buttons on the site (SubscribedButtonClick). Through Meta's automatic matching it may also receive hashed identifiers. It does not receive anything you enter in the app. Meta's handling of this data is described in its privacy policy.

When you first visit we ask you to choose: “Accept all”, “Reject non-essential”, or pick by category under “Settings”. We store your choice with its date and ask again after 12 months or if this policy changes. You can change or withdraw consent any time with the “Cookie settings” link in the footer of every page.

Premium subscriptions are sold and billed by Apple (App Store) or Google (Google Play). We never see your card details. Manage or cancel a subscription from your App Store or Google Play subscriptions.

Kidsit AI is for parents and guardians aged 18 or over, not for children. Any information about a child is entered by their parent or guardian and is handled as described above. The digest is calculated on the device for each child and is not sent unless you turn on sharing and confirm a send. Anonymous insights are not meant to include a child's name or other identifiers.

Under the Israeli Privacy Protection Law and the GDPR, you have the right to access, correct, or delete your data. On-device data is under your control in the app. Anonymous insights are not linked to an identity, so we cannot look up a specific send; free text is deleted after 90 days. Contact us using the address below to exercise these rights.

We use HTTPS and standard security measures to protect information in transit and on our site.