How we handle your inner world.

Who we are

All My Parts (“the app,” “we”) is an AI-assisted journaling product built around ideas from Internal Family Systems. It is not therapy. The service is operated by Studio Maya Inc., an Ontario-based corporation. It is for adults (18 and older) in the United States and Canada.

Our privacy officer is Humayun Khan, founder. For privacy questions, requests, or complaints, contact help@allmyparts.me.

What we collect

• Account info. Your email address and the basic profile details that come with signing in (such as your name), through our authentication provider, Clerk. If you sign in with Apple or Google, we receive only the name, email address, and (for Google) profile picture they share with us, never your password. Apple may give us a private relay email address instead of your real one.
• About you. Answers you choose to give during onboarding and in Settings: a display name, age range, gender, how familiar you are with IFS, and what brings you here (for example “anxiety” or “old wounds, trauma work”). On the website you can also add a date of birth to your profile.
• Your journal content. Everything you write in the app: entries, your conversations with the journaling AI (your messages and its replies), parts you name, memories, things you ask the app to remember, and images you add to parts.
• Voice. If you use voice input (a Pro feature), the recording is sent to OpenAI to turn it into text. We do not keep the audio on our servers, only the text if you add it to an entry.
• What the AI writes about you. After a session, the AI writes a title and a reflection, picks out feelings, people, and topics, and suggests parts. Over time it adds detail to each part (such as its purpose, fears, beliefs, triggers, and where you feel it in your body) and builds a self-profile for you. It also writes weekly reports. This can reveal sensitive things about your emotional and mental health. See “The AI-built profile” below.
• Condition ratings. If you use the optional Conditions screen on the website, the ratings you give yourself.
• Feedback and reports. If you send feedback or report an AI response, we keep what you send, along with your email address, your account ID, and the page you were on.
• Safety counts. Each time our safety check shows you a “Get help now” card, we add one to a running count. The count records only the date, the level and type of concern the check found, which part of the app showed the card, the platform (iOS, Android, or web), and whether the account was free or Pro. It does not include your words and does not identify you.
• Usage signals. Timestamps, which features you use, error logs, and basic technical details (such as your IP address, device or browser type, and app version) that our hosting and service providers log when you use the app. For each AI request we also log which feature made it, how many tokens it used, and its estimated cost, tied to your account ID. Those logs never include your content. On the website we use Vercel Web Analytics to count page views. It does not use cookies and does not receive what you write. We do not use advertising trackers.
• Crash reports. If the mobile app hits an error or crashes, it sends us a short technical report: the error message, where in the app's code it happened, the app version, and your device's platform and system version. The report does not include your account ID or anything you write, and it is not linked to you. We use it only to find and fix bugs. Apple and Google also share crash statistics with us from people who turned on sharing analytics with app developers in their device settings.
• Payment info. On the web, card details are handled by Stripe. In the apps, payment is handled by Apple or Google. We store only your subscription status and the matching IDs from Stripe or RevenueCat. We never see your card number.

How we use it

• To run the journaling features: the AI conversation, the reflection after each session, your parts and profile, and weekly reports.
• To help the AI remember you. On Pro accounts, the AI also receives your main parts, your self-profile, facts you asked it to remember, and snippets of similar past entries, so it can refer back to them.
• To power search. To make “Ask your journal” (Pro) and recall of past entries work, we send your entry text to OpenAI to create embeddings (a numeric fingerprint of the text) and store them with your entries.
• To offer voice input and read-aloud (both Pro).
• To look for signs of risk in your messages and show you help (see “Safety and crisis content” below).
• To handle billing, answer your messages, keep the service secure, and prevent abuse. We do not send marketing emails without explicit opt-in.

What goes to our AI providers. We send your text to Anthropic (Claude) for the live conversation, the reflection and extraction after a session, building your parts and profile, the suggested prompts on your dashboard, weekly reports, Ask answers, and the AI checks in our safety protocol. We send text to OpenAI for search embeddings, and audio and replies for voice input and read-aloud.

Names. For the reflection and extraction after a session, weekly reports, and Ask answers, we replace the names of people we already know from your entries with placeholders before sending, then put the names back in the result. This only catches names we already know. It does not catch every name or identifying detail. Everything else we send to AI providers, including the live conversation, building your parts and profile, search embeddings, voice input, and read-aloud, goes as written, because the AI needs the names to respond naturally.

What we don't do

• We do not sell your data. Ever, to anyone.
• We do not use your journal content to train AI models, ours or anyone else's. Under the terms we use with Anthropic and OpenAI, content sent through their APIs is not used to train their models.
• We do not show ads and do not work with ad-tech vendors or data brokers.
• We do not contact anyone on your behalf. We do not notify your emergency contacts, emergency services, or anyone else, even when our safety check shows you help.

The AI-built profile

The app builds a picture of you over time from what you write. An AI model reads your entries and writes the notes, with no person reviewing them. The picture includes your parts and a self-profile covering your core beliefs, strengths, values, growth edges, emotional and relational patterns, coping strategies, fears, “wounds” (unresolved pain), and self-awareness. We use it to personalize the app and, on Pro accounts, to help the AI remember you between sessions. We do not use it to make decisions about you outside the app.

You can view the profile on the website, export it, and edit or delete your parts. You can remove everything by deleting your account. In the mobile apps, you can turn AI features off in Settings.

We used to ask the AI to look for signs of five conditions in your entries: autism, body dysmorphia, OCD, complex PTSD, and borderline personality disorder. We have stopped doing this for everyone. Notes it saved earlier are hidden in the app, and they are removed when you delete your account. The ratings you give yourself on the Conditions screen are yours and stay.

How we protect it

• Encryption in transit. All traffic to and from the app uses TLS.
• Encryption at rest. Most of the text you write and most of what the AI writes about you is encrypted at the application layer with AES-256-GCM before it is stored, using a key managed separately from the database. That covers your messages and the AI's replies, entry titles, reflections, part names and details, the names of people you mention, memory names and descriptions, your self-profile, the notes you ask the app to remember, the text of weekly reports, and the text kept for search. Some things are not encrypted this way: your account and onboarding details, topic and mood tags, report titles and counts, your condition ratings, feedback you send us, and the numeric search fingerprints. These stay behind the encrypted connections and database controls described here.
• Isolation enforced by the database itself. Requests from signed-in users run under a restricted database role with row-level security: the database refuses to return rows that don't belong to the signed-in user. Your data being separate from other users' doesn't depend on application code alone.
• What we can and can't see. Day-to-day operations never involve reading anyone's journal. Direct database access is restricted, and most content is ciphertext at rest, with the decryption key stored separately. To be fully honest: like any hosted service that processes your content with AI, our servers must briefly hold entries in readable form in memory while generating a response, so this is not end-to-end encryption. The protections above are designed to make access outside that moment as hard as possible, and never routine.
• Authenticated uploads. Images you add are stored in a private Cloudflare R2 bucket and served only through an authenticated proxy. The raw URL is not enough to retrieve them.

Who we share with

We use the service providers below to run the app. We share data with them only for that purpose, and each one's own terms and privacy commitments limit how it can use what we send.

• Anthropic: Claude, which writes AI replies and reflections, builds your parts and profile, writes weekly reports, answers Ask questions, and runs the AI checks in our safety protocol.
• OpenAI: creates search embeddings of your entries and questions, transcribes voice recordings (Pro), and turns AI replies into speech (Pro).
• Clerk: sign-in and account management. Holds your email address, name, sign-in method, and the record of the consents you give.
• Neon: PostgreSQL database hosting (US).
• Vercel: application hosting, edge network, server logs, and web analytics.
• Cloudflare: private image storage (R2) and bot protection (Turnstile) at sign-up.
• Stripe: web payments and subscriptions. Receives your email address, account ID, and payment details.
• RevenueCat, Apple, and Google: App Store and Google Play subscriptions. Apple and Google process the payment, and RevenueCat tells us your subscription status. Apple also provides Sign in with Apple.
• Slack: our team chat. Feedback and AI-response reports you send are posted there with your email address and account ID.
• Expo: delivers updates to the mobile app. Sees basic request details, such as your IP address and app version, when the app checks for an update.

These providers may keep copies for a limited time under their own policies (for example, to detect abuse) before deleting them. We may also disclose information if the law requires it.

Where your data is processed

Your data is stored and processed in the United States, by us and by the providers above. If you live in Canada, that means your information is transferred outside Canada and can be accessed under US law. By using All My Parts you consent to your data being transferred to and processed in the United States, subject to contractual safeguards.

How long we keep it

We keep your content until you delete it or delete your account. When you delete your account, we erase your data from our active systems right away, and copies in encrypted backups are purged within 30 days. Some records stay: payment providers (Stripe, Apple, Google, RevenueCat) keep their own transaction records as the law requires, our AI usage logs (counts and costs, no content) are kept, and feedback posted to Slack may remain there. Crash reports are deleted after 90 days.

Your rights

Wherever you are, you can export your journal, parts, memories, profile, and reports as a single Markdown file. On the website, go to Settings → Data. In the mobile app, go to Settings → Download your entries. For anything else we hold about you, email us.

You can delete entries, parts, and memories from inside the app. To have the facts you asked the app to remember removed without deleting your account, email us.

To delete your account and everything in it, open the mobile app, go to Settings, and tap Delete account. It takes effect right away. If you only use the website or can't open the app, email help@allmyparts.me from the address on your account and we'll delete it within 30 days. More detail is on our account deletion page. Deleting your account does not cancel a paid subscription. Cancel it first in your App Store or Google Play settings, or on the website if you subscribed there.

If you are in Canada, you have rights under the Personal Information Protection and Electronic Documents Act (PIPEDA) and applicable provincial privacy laws, including the right to access, correct, and request deletion of your data, and to file a complaint with the Office of the Privacy Commissioner of Canada.

If you are in the European Economic Area or United Kingdom, you have rights under the GDPR including access, rectification, erasure, restriction, portability, and objection. We are limiting availability of the app to the United States and Canada.

If you are in California, you have rights under the CCPA/CPRA including access, deletion, and the right to opt out of any “sale” or “sharing” of personal information. We don't do either.

Consumer health data

Some of what you write or tell us may be information about your mental or emotional health. Washington, Nevada, and Connecticut give you specific rights over this kind of information. Our Consumer Health Data Privacy Policy explains what we collect, why, who we share it with, how consent works, and how to use your rights and appeal.

Safety and crisis content

All My Parts is not a crisis service, and no one at All My Parts reads your messages as you write them. We built some automatic protections, and the full protocol is on our Safety & Crisis Protocol page. In short:

• The AI is instructed to ask you directly about safety and offer 988 at any sign of suicidal thinking or self-harm, not only when danger is imminent. It never gives information about methods or locations, and it points deep trauma work, and any suicidal or self-harming parts, to a therapist.
• An automatic check looks at each message you send, using a keyword screen and a second AI check. When it finds signs of risk, the app shows a “Get help now” card with 988 (call or text), 911, and findahelpline.com. No person sees your message as part of this.
• After a session that involved a crisis, the summary gets a plain title, no quote, and a line pointing to support, and the app does not create part cards from it.
• We count how often the card is shown, without your words or anything that identifies you (see “Safety counts” above).

Cookies and local storage

We use only the cookies that are essential to the service: authentication and session cookies set by Clerk. We do not use analytics, advertising, or social tracking cookies. On the website, your browser also keeps an unsent draft of what you are writing in local storage so you don't lose it. That draft stays on your device.

Children's privacy

All My Parts is for adults. You must be at least 18 to use it, and we do not knowingly collect information from anyone under 18. If you are under 18, please do not create an account. If we learn that we have collected information from someone under 18, we will delete it.

Changes to this policy

We'll update this page when we change anything material. If the change affects how we use your existing data, we'll email you before the change takes effect.

Contact

Our privacy officer is Humayun Khan, founder. Email help@allmyparts.me for anything privacy-related: questions, requests, complaints. We'll respond within a week, usually sooner.