AI and your data

Does The Peachy Life send my journal to AI?

Yes, some of what you write is sent to an AI company when you use an AI feature. The Peachy Life sends that text to Anthropic in the United States. OpenAI reads any food photo you choose to analyse, and is a standby we can switch some non-chat features to. We don’t attach your name, email or account ID, though anything you type, including names, reaches the provider as you wrote it. We don’t use your data to train any AI model ourselves, and AI features are optional: nothing is sent until you accept them.

This page exists because our Privacy Policy is legally complete, and long. Section 5a already says everything below, but you should not have to read the whole policy to find out where a sentence you wrote at 11pm ends up.

AT A GLANCE

The short version

Who: Anthropic (default, and the only provider that receives your chat) and OpenAI (food photos, and a standby we can switch some non-chat features to). Dictation goes to Apple or Google. Food searches go to the USDA.
Where: All of those companies process in the United States. Your journal, mood and chat records are stored in Google Cloud’s Sydney region.
Training: We don’t use your data to train or fine-tune any AI model ourselves. The providers’ own positions are quoted below.
Optional: Nothing goes to an AI provider until you accept AI features. Mood tracking, journalling and assessments work without them.

FEATURE BY FEATURE

What leaves your phone, feature by feature

Each row is one thing you can do in the app. All of these providers process in the United States.

FEATURE	WHAT IS SENT	TO WHOM	HOW TO AVOID IT
AI chat	The text of your message, plus a rolling summary of the recent conversation	Anthropic only. Chat never goes to OpenAI	Don’t use the chat. Nothing is sent until you send a message.
Weekly insight	Your recent check-ins, journal themes, and habit, exercise and food context	Anthropic	Written only if you accept AI features. You can turn off individual data sources in Settings.
Journal analysis	The text of the entry you asked to have analysed	Anthropic	Don’t tap analyse. Plain journalling sends nothing to an AI provider.
Assessment report	A descriptive band only, such as “a developing area”. Never your score or your answers	Anthropic; OpenAI only if we switch this feature to our standby	Skip the written report. The assessment still works.
Wearable summaries	Summaries derived from your wearable data, such as sleep, heart-rate variability, activity and engagement patterns	Anthropic	On by default when you connect a wearable. Turn it off in Settings (see below).
Photo food logging	The photograph	OpenAI	Log food by searching or typing instead. There is no on-device option.
Meal description	The food text you type when you describe a meal	Our AI providers, as set out in Privacy Policy section 5a	Search the food database or enter the food manually instead.
Dictation	The audio of whatever you say	Apple (iPhone) or Google (Android), through your phone’s speech recogniser	Type instead. Nothing is sent unless you tap the microphone.
Food search	The search term you type	USDA FoodData Central, a US government agency (not an AI provider)	Enter the food manually.

Before any text goes to Anthropic or OpenAI, our server scans it for email addresses and phone numbers and strips the ones it finds. That is pattern matching. It misses things, and it cannot remove a name, a place, or what you wrote about someone else. Assume anything you send for analysis reaches the provider substantially as you wrote it.

WORTH KNOWING

The one setting that is on by default

When you connect a wearable, the setting that lets the AI use summaries of your wearable data is switched on at the same time, and it stays on until you turn it off. Everything else in the table above is off until you use it. We are saying so plainly because it is the opposite of how an opt-in setting works. To turn it off, open Settings in the app and switch off the option that lets the AI use your wearable data. Your wearable keeps syncing; the AI simply stops being told about it.

One honest limitation: once you have accepted AI features, the app lets you narrow what is shared (turn off individual data sources, delete your chat history, and clear what Peachy remembers, in Settings > Privacy & Data > Data Controls & Export), but a single switch that turns every AI feature back off is not built yet. You can withdraw your consent at any time; until that switch exists, the way to fully stop AI processing is to delete your account, which removes your data apart from the records the Privacy Policy says we keep after deletion.

What we never send to an AI provider

Your account ID, your email address, your display name, your numeric assessment scores and your answers to individual assessment questions. A request reaches Anthropic or OpenAI as text with nothing that links it back to your account.

Is my data used to train AI?

We don’t use your data to train or fine-tune any AI model. We send text to Anthropic and OpenAI under their standard commercial API terms for business customers. Here is what each of them publishes about that data, checked on 27 September 2026:

“Retained data is never used for model training without your express permission.”

On retention: “Conversation content (your prompts and Claude’s outputs) is not retained by default”, with exceptions: some models require 30-day retention, and content flagged by its safety systems may be kept “for up to 2 years”.
Anthropic, API and data retention

“As of March 1, 2023, data sent to the OpenAI API is not used to train or improve OpenAI models (unless you explicitly opt in to share data with us).”

We have not opted in. On retention: abuse-monitoring logs are “retained for up to 30 days, unless longer retention is required by law”.
OpenAI, Data controls in the OpenAI platform

These are the providers’ public statements, not our guarantee. We do not yet hold written confirmation from Anthropic about how it handles the data we send through its API. When that changes, this page and section 5a of the Privacy Policy will change with it.

Three things that get confused

No training. The provider does not use your text to improve its models. Both providers publish this for business API traffic, quoted above.
No retention. The provider does not keep your text after replying. Neither provider’s default is zero in every case, and we have not arranged a zero-retention tier.
On-device. The text never leaves your phone. The Peachy Life does not work this way: every AI feature is processed on a provider’s servers.

Who can read it?

Your journal is not end-to-end encrypted. That would make the AI features impossible, because our server has to read your text to strip contact details and send it to a provider. Instead:

In transit and at rest: everything moves over encrypted connections, and your records are encrypted at rest.
Where: your journal entries, mood check-ins, chat history, assessments and wearable summaries are stored in Google Cloud’s Sydney region. We moved our primary database there on 2 August 2026. Some processing still runs on servers in the United States, including for people on older versions of the app.
Who: within our systems, only you and our authorised engineers can read your data. Engineers are bound by confidentiality obligations, and every access is audit-logged.
How long: your records are kept for as long as your account exists and erased as soon as you delete it. AI chat messages are kept for up to 90 days from each message; the automated deletion job for them is still being implemented, so until it is live they may be kept longer.

FAQ

Questions people ask

Does The Peachy Life train AI on my journal entries?

We don't use your data to train or fine-tune any AI model. Our AI providers publish their own positions for business API traffic: OpenAI says API data is not used to train its models unless the customer opts in, and we have not; Anthropic says retained data is never used for model training without express permission. Those are their statements, not our guarantee, and we do not yet hold written confirmation from Anthropic about our traffic.

Is it OpenAI?

Mostly not. Anthropic, in the United States, is the default provider and the only one that ever receives your chat. OpenAI, also in the United States, handles food photos, and is a standby we can switch some non-chat features to if Anthropic is unavailable.

Does the AI run on my phone?

No. Nothing in The Peachy Life runs an AI model on your phone. Every AI feature sends text or a photo to a provider's servers in the United States and gets a reply back. Dictation is also processed off the device, by Apple or Google.

Can I use The Peachy Life without the AI?

Yes. Nothing is sent to an AI provider until you accept AI features, and mood check-ins, journalling and assessments all work without them. Once accepted, you can narrow what is shared in Settings, but a single switch that turns AI features back off is not built yet; until it is, the way to fully stop AI processing is to delete your account.

Is my journal end-to-end encrypted?

No. It is encrypted in transit and at rest, and stored in Google Cloud's Sydney region. End-to-end encryption would stop the AI features working, because our server has to read the text to strip contact details and send it to a provider.

What happens if I dictate instead of typing?

Your audio goes to Apple (iPhone) or Google (Android) in the United States to be turned into text, through your phone's built-in speech recogniser. If you would rather your voice was not sent anywhere, type instead: nothing is sent for transcription unless you use the microphone.

Does anyone at The Peachy Life read my journal?

Only you and our authorised engineers can read your data within our systems. Engineers are bound by confidentiality obligations and every access is audit-logged.

YOUR CALL

If you change your mind

You can leave with everything, any time, without asking us. In the app, Settings > Export Data gives you a machine-readable copy of your records (the download link expires after 24 hours). You can delete your account from Settings → Account & Security → Manage account deletion; the account deletion page explains exactly what happens and when.

The full legal version of everything on this page is Privacy Policy section 5a. If anything here does not match what the app actually does, tell us and we will fix the page or the app.

If you need someone right now, crisis support is one tap away in the app and on this site, and it is never behind a paywall.