MorseSign inStart freeOpen Morse

Is Claude private for your health records? Connectors, Apple Health, and HIPAA

The Morse teamPublished 9 min read

As of November 2026, Claude's health features keep connected records out of model training and out of memory by default, but they aren't private from Anthropic, and HIPAA doesn't follow your records into Claude. Anthropic says, "We do not use users' health data to train models" (Anthropic). Its HealthEx guide says chats that use the HealthEx integration "are never used to train our models" (Anthropic). Anthropic edited its launch post on February 7, 2026 "to clarify that HIPAA-ready products are specific to healthcare providers and payers" (Anthropic). Anthropic sells those products as Claude for Healthcare, for "healthcare providers, payers, and health tech companies," separately from the consumer health features (Anthropic). Anthropic's pages describe health "integrations" and connectors rather than a product named Claude Health, so the phrase people search covers three things: record connectors, Apple Health and Android Health Connect access, and health questions typed into an ordinary chat.

This isn't medical advice. We read Anthropic's launch post, its iOS, Android, and memory help pages, its HealthEx guide, its training and retention pages, and its Consumer Health Data Privacy Policy on November 15, 2026, along with the connector pages and privacy policies of HealthEx, Function, and athenahealth and an HHS guidance page. Anthropic earns credit on memory. Claude doesn't save health details to memory unless you turn on a setting for sensitive topics, and turning it off again removes what was saved.

SettingDefaultRetentionSource
HealthEx, Function, or athenahealth connectorOff until you connect it; US only, on paid plansHealthEx consent lasts up to 1 yearAnthropic
Apple Health or Health Connect accessOff until you allow each data type; read-onlyClaude's tool use is kept in chat historyAnthropic
Training on chats that use HealthExNever, whatever your training settingNot applicableAnthropic
Training on health questions you type or pasteYour "Help Improve our AI models" setting; no health exception listedUp to 5 years, de-identified, if the setting is onAnthropic
Health details in memoryOff until you turn on "Include sensitive topics in memory"Removed when you turn the setting offAnthropic
Chats you deleteKept until you delete themDeleted from back-end storage within 30 daysAnthropic
HIPAA-ready productsProviders and payers onlyNot applicableAnthropic

What health data can you connect to Claude?

Medical records, lab results, and phone health data, if you're in the US on a paid plan. Anthropic's January 11, 2026 launch post said "Claude Pro and Max plan subscribers can choose to give Claude secure access to their lab results and health records," through HealthEx and Function connectors, with Apple Health and Android Health Connect in the mobile apps (Anthropic). Anthropic's iOS help page now says the health features "are currently in beta, require a Pro, Max, or Ultra plan, and are available in the US only" (Anthropic).

  • HealthEx pulls records from your providers' patient portals. Its tools in Claude include medications, lab results, conditions, allergies, and care plans (Anthropic). HealthEx verifies you with CLEAR, which usually takes a phone number, a selfie, and photos of a US government ID, and Anthropic says HealthEx "doesn't share biometric data or ID information with Claude" (Anthropic).
  • Function shares a summary of a Function member's lab results and nutrition plan. Its own setup page says that once connected, "Claude can access the information you've authorized to be shared from any conversation" (Function).
  • athenahealth (beta), added in September 2026, links Claude to your athenahealth medical record, including encounter notes and lab results (Anthropic).
  • Apple Health gives Claude activity, workouts, vitals, body measurements, sleep, and nutrition, and Claude "can read health data but cannot write or modify entries" (Anthropic).
  • Android Health Connect needs Android 14 or later, and Claude "can see data from any app you've connected to Health Connect" (Anthropic).

HealthEx's own guide says its records can include "your HIV status, STIs, substance use, reproductive health, disabilities, and other protected health information" (HealthEx). Uncheck what you don't need on the consent screens. Anthropic's guide says data sets on Epic sites "are checked by default" and you can unselect any of them (Anthropic).

Does Claude train on your health data?

Not on data that comes through the health connectors, but a health question you type yourself follows your normal training setting. The HealthEx guide says "Chats that use the HealthEx integration are never used to train our models" (Anthropic). Anthropic's training page for Free, Pro, and Max says raw content from connectors isn't used, "though data may be included if it's directly copied into your conversation with Claude" (Anthropic). That page names no exception for health.

So a lab report you paste into a regular chat is treated like any other chat. If "Help Improve our AI models" is on, Anthropic may keep it "in a de-identified format for up to 5 years" (Anthropic). Turn it off under Settings, then Privacy (Anthropic), or use an incognito chat, which the training page says isn't used "even if you have enabled Model Improvement." Chats "flagged for safety review" can still go into models for Anthropic's Safeguards team (Anthropic), as we cover in is Claude private.

Anthropic's Consumer Health Data Privacy Policy, effective January 12, 2026 and written for Washington and states with similar laws, lists one more use: "In an aggregated form, to help us develop, test and improve our services" (Anthropic). It doesn't define the aggregated form.

Does Claude remember your health information?

Not by default. Anthropic's memory page says Claude "does not store topics related to personal or sensitive subject matter, like your health" (Anthropic). The switch is Include sensitive topics in memory, under Settings, then Your Claude, and Claude offers it in a one-time notice the first time it declines to save something sensitive. Decline the notice, or turn the setting off later, and "Claude removes any sensitive items already saved to memory." The HealthEx guide adds that Claude is designed not to use sensitive health data in future conversations, "while still keeping in mind your areas of concern" (Anthropic).

Is Claude HIPAA compliant for your own records?

No. HHS says that once health information reaches, at your direction, an app that is neither a covered entity nor a business associate, "the information is no longer subject to the protections of the HIPAA Rules" (HHS). Anthropic's Consumer Health Data Privacy Policy doesn't mention HIPAA (Anthropic). HealthEx, Function, and athenahealth each describe their own HIPAA role, in the table below. Anthropic's BAA page covers its commercial products, such as Claude for Work and the API, and sends Free, Pro, Max, and Ultra to a separate consumer page (Anthropic). The business side is in HIPAA compliant AI chatbots.

Where does each copy of your records end up?

In up to three places, each under different rules. HealthEx stores your records on its own servers, under a deletion clock separate from Claude's.

CopyWho holds itRules it says applyHow long it stays
Your chart at the clinicYour provider and its vendorsHIPAA; athenahealth describes its role as "a Business Associate to your healthcare provider" (athenahealth)As long as its contracts with your provider allow
Records HealthEx retrievesHealthExIt "follows the HIPAA privacy rule" (HealthEx), but acts as "an independent 'data controller'" and not a "business associate" of the apps you share with (HealthEx)While consent is valid, then deleted within 30 days; backups up to 180 days
Function lab summaryFunctionIt "may operate as a 'business associate'" but "HIPAA does not apply to all personal information that we process" (Function)Not stated for the connector
What Claude reads into a chatAnthropicAnthropic's Privacy Policy and Consumer Health Data Privacy Policy (Anthropic)Until you delete the chat, then within 30 days

HealthEx's policy, effective November 12, 2026, says that once you share health data with an app, "that party becomes responsible for how they use it" (HealthEx). Withdrawing consent at HealthEx "will not affect data that has already been shared with third parties." Anthropic's iOS page says that when Claude uses a phone tool, "their use is preserved in the conversation history in the same way as any tool" (Anthropic). The copy in the chat lasts until you delete it.

How do you disconnect your health data and delete it?

Take four steps, because each copy has its own control.

  1. Disconnect the connector in Claude. Open Settings, then Connectors, and disconnect HealthEx or Function (HealthEx).
  2. Withdraw consent at HealthEx, or delete your HealthEx account. Disconnecting in Claude doesn't do this for you. HealthEx deletes an account's data within 24 hours, except audit and backup logs, and keeps backups for up to 180 days (HealthEx).
  3. Revoke phone access. On iPhone, go to Settings, Health, Data Access & Devices, then Claude (Anthropic). On Android, go to Settings, Apps, Health Connect, App permissions, then Claude (Anthropic).
  4. Delete the chats that used your records. A deleted chat leaves your history at once and Anthropic's back-end storage within 30 days (Anthropic). Our guide on how to delete Claude chats has the steps.

Residents of Washington and similar states may also ask Anthropic to delete the consumer health data it collected, and the policy points to Claude's Privacy Settings for its data tools (Anthropic).

Is Claude more private than ChatGPT Health?

Neither is clearly ahead, because they're built differently. OpenAI syncs connected records into its own systems, keeps them out of training "regardless of the model-training setting you choose," and deletes them "within 30 days" of disconnecting (OpenAI), as our page on whether ChatGPT Health is private covers. Claude pulls only what a question needs from a partner like HealthEx, which keeps the full copy under its own policy (Anthropic).

How Morse handles a lab result you want explained

Every protection above is a promise attached to a stored copy: the one HealthEx holds, the chat Anthropic keeps under your account, and the training rule that changes with how the data arrived. Morse's servers don't keep your messages. Your chat history is stored in your browser, encrypted, not on Morse's servers, and Morse doesn't train AI on your chats, sell your data, or show ads (privacy policy).

  • E2EE, for a diagnosis or a lab panel you'd show no company. Your browser checks the enclave's hardware attestation before anything leaves your device, then encrypts the message so only that verified enclave can read it. Morse's servers relay ciphertext they can't read. Third-party providers run the enclaves, not Morse, and the protection rests on the enclave hardware and attestation working as designed. E2EE uses open-weight models such as GLM, Kimi, and DeepSeek, with no web search or image generation, so paste the values and reference ranges you want explained.
  • Anonymous, when you want Claude itself to read the result. Morse's proxy removes your account, email, and IP address and sends the question under Morse's own account, so the provider sees the question without knowing it's you, and may keep it for a limited time to check for abuse, under its own policy. Cut your name, birth date, and record numbers before you paste.
  • Private, the default, for questions about sleep, steps, or a side effect. Open-weight models and Grok answer with your identity removed. Morse has zero data retention agreements with its inference, search, and other providers, so they've agreed to keep nothing, and where a provider confirms zero retention on each reply, Morse checks it and refuses a reply that doesn't.

Because we don't store chats, we usually have very little to give authorities. Clearing your browser or losing the device deletes those chats, and Morse can't recover them, so save the questions you want to bring to your appointment.

So is Claude private enough for your health records?

  • Model training: connected records are excluded, but pasted results follow "Help Improve our AI models," so turn it off or use incognito. Morse doesn't train AI on your chats.
  • Memory: leave Include sensitive topics in memory off.
  • A partner's copy: withdraw consent at HealthEx when you're done, not only in Claude.
  • Anything you'd keep from every company: start the chat in E2EE on Morse.

You can start free or send a first message from the Morse home page with no account. E2EE chats are part of Pro, at $19 a month.

Common questions

Can I use Claude's health features outside the US?

Not as of November 2026. Anthropic's iOS help page says the health features are in beta and limited to users in the US, and the HealthEx connector is offered to Claude users in the United States.

Do I have to show ID to connect my medical records to Claude?

Through HealthEx, yes. HealthEx checks your identity with CLEAR, which usually means a selfie and photos of a US government ID, and Anthropic says HealthEx doesn't share that biometric or ID information with Claude.

Can Claude change my Apple Health data?

No. Anthropic's help pages say Claude can read health data but can't write or modify entries in Apple Health.

Is Claude for Healthcare the same thing as Claude's health features?

No. Claude for Healthcare is Anthropic's set of HIPAA-ready products for providers and payers. The personal health connectors and Apple Health access are consumer features of the Pro, Max, and Ultra plans.

Ask what you actually want to ask.

80+ models, characters, and images, with every chat encrypted and kept on your device. Free to start.

Start a private chat

Read next