What the Anthropic API means for your app's privacy disclosures

the Anthropic API is a model API. What it involves, whether it needs the tracking prompt, what goes on the App Store labels and the Play Data safety form, and the rejection it is usually behind.

Updated 2026-09-18 · Written from the same rules terms.gg uses to generate documents

What the Anthropic API actually does

The same shape of integration, with Claude behind it.

What its job necessarily involves

These follow from what the thing is for, so they are true whichever version you installed:

• User input and the context your application adds to it.
• Nothing retained for training on the commercial API, which is worth stating in your policy because people ask.

What to open and check in the Anthropic API

The settings below decide what the honest answer on the forms is, and they are specific to this SDK rather than to its category:

• Whether calls go from the device or from your server, for the same key exposure reason.
• What your system prompt carries about the user.

the Anthropic API does not need the tracking prompt

Nothing here reads an advertising identifier. So the Anthropic API does not on its own require App Tracking Transparency. That is a statement about this SDK, not about your app: add one ad network or one attribution tool and the prompt applies to everything.

the Anthropic API in Europe: runs on necessity

You cannot deliver the service without it, so the Anthropic API generally rests on the contract with your user rather than on consent. Say plainly in the policy that it is a processor, what it does, and where it runs.

the Anthropic API in California: probably not a sale

the Anthropic API does its job for you rather than for its own commercial purposes, which normally keeps it a service provider rather than a sale. That depends on your contract saying so, and the standard terms usually do.

the Anthropic API on the App Store privacy labels

Apple asks what your app collects, and your app collects whatever its SDKs collect. the Anthropic API is generally declared under App Functionality, alongside whatever the rest of your app does for its own reasons.

the Anthropic API on the Google Play Data safety form

The form asks two things the labels do not: whether data is shared with anyone else, and whether it is encrypted in transit. For this SDK the honest answer is usually collected but not shared, because it processes on your behalf.

Check Anthropic's own privacy manifest for the Anthropic API

Since 2024 Apple has required third-party SDKs on its list to ship a signed privacy manifest declaring what they collect and which sensitive APIs they use, and your app's combined manifest is built from them. That file is the authoritative answer for the Anthropic API, it comes from Anthropic, and it changes when they ship.

The rejection the Anthropic API is usually behind

Users were never told their messages reach a model provider, and a support transcript is exactly the kind of content that needs saying.

What to put in the privacy policy about the Anthropic API

• Anthropic named as a processor or a recipient, not hidden behind "our partners".
• What it is for, in a sentence someone would understand without knowing what an SDK is.
• Where it runs, and what makes a transfer outside your region lawful.
• How long the data is kept, which is a question the vendor documents and most policies skip.
• That it processes on your instructions rather than for its own purposes.

Keeping it true after launch

A legal page stops being true the moment the product moves past it, usually by adding a payment provider, an analytics SDK or a sign-in. No store re-checks your pages against your build, so the drift is yours to notice.

• Re-read it whenever you add a dependency that sees user data.
• Re-check what loads on the page after any change: the cookie notice and the policy have to agree.
• Keep the URL stable. Changing where a policy lives breaks every listing that points at it.