Detailed Analysis
A Claude Pro subscriber reports on Reddit that the Apple Health connector has vanished from their available integrations within the Claude app, leaving only Google, Canva, and HubSpot listed. The user had been actively relying on the connector for over two weeks to track sleep data and conduct morning health check-ins — including, notably, a session the previous day using Claude's Opus 4.7 model. Despite confirming that iOS permissions remain fully granted on the device side, the connector no longer appears on Claude's end, and the user has found no official communication from Anthropic acknowledging a change or regression.
The research context is notable for what it does *not* contain: no search results confirm that Anthropic ever officially launched or documented a Claude-native Apple Health Connector. This raises the possibility that the feature existed as a limited, experimental, or beta integration that may have been quietly rolled back rather than formally deprecated. The user's experience — two-plus weeks of reliable functionality followed by a sudden disappearance — is consistent with a server-side or backend change on Anthropic's infrastructure rather than a client-side iOS permissions issue. The iOS settings still show the connection as active, which rules out a user-error or device-level explanation and points squarely toward a change on Anthropic's platform.
This incident fits into a broader pattern in the AI assistant space where health data integrations are emerging as a competitive frontier. OpenAI, for instance, was reported in December 2025 to be developing an Apple Health Connector for ChatGPT, with hidden code revealing plans for syncing activity, sleep, diet, and other biometric data. If Anthropic had been testing a similar capability, its apparent withdrawal — without user notification — reflects a recurring friction point in how AI companies manage experimental feature rollouts: users adopt and become dependent on features before those features are production-stable, and quiet removals create trust erosion.
The broader significance lies in how personal health data integration changes the nature of AI assistant relationships. The user explicitly frames their reliance on Claude not as casual usage but as support through "a significant shift" in their life, with structured morning check-ins serving a quasi-therapeutic or wellness-coaching function. When infrastructure supporting that kind of continuity disappears without explanation, the impact extends beyond inconvenience into a disruption of an established routine built around AI assistance. This underscores that as AI companies expand into sensitive personal data domains like health tracking, the bar for reliability, transparency, and communication around feature changes must be substantially higher than for general-purpose tools.
Read original article →