Legal disclaimers

Last updated: 2026-09-07

This page is general, good-faith boilerplate, not a legal opinion -- it isn't written or reviewed by a lawyer, and it isn't a substitute for one. If anything here is compliance-critical for your own use of this site or this library, consult your own counsel. Nothing on this page should be read as a guarantee of any particular legal, accessibility, or privacy-compliance status.

For the library's and this site's copyright licensing (MIT, the vendored Apache-2.0 uap-core data, and the CC0-1.0 logo icons), see License -- this page doesn't repeat that content.

Trademarks & third-party names

kmp-user-agent parses and generates User-Agent strings, which means its detection logic references third-party browser, operating system, device, bot, and AI-agent crawler names and identifiers (for example "Chrome", "Windows", "iPhone", "Googlebot", "GPTBot"). Every such reference is nominative use: identifying which real-world product or service a given User-Agent string matches. It does not imply affiliation with, sponsorship by, or endorsement from any of the companies or projects named. All product and company names are trademarks of their respective owners.

The site's homepage also displays third-party brand logo icons (via Simple Icons) next to some of these detection results, for the same identifying purpose. See License § Third-party attribution: docs-site logos for the specifics of that attribution and the same nominative-use framing.

Accessibility

This documentation site is built on VitePress's default theme, used largely as-is. No formal accessibility audit and no WCAG conformance testing (at any level -- A, AA, or AAA) has been performed against this site, and we make no claim of WCAG or other accessibility conformance. That said, specific interactive elements were built with accessibility in mind on a best-effort basis: for example, the analytics consent banner uses an ARIA dialog role with aria-live and moves keyboard focus to itself when it appears, so it doesn't render silently outside a keyboard or screen-reader visitor's attention. This is targeted, uncertified engineering care for that one component -- not a substitute for an audit, and not a claim that the rest of the site has received the same scrutiny.

If you encounter an accessibility barrier using this site, please open an issue so it can be looked at. The same issue tracker is also the right place for privacy questions -- see the Privacy & analytics section below.

Privacy & analytics

This site uses Google Analytics (GA4), gated behind a cookie-consent banner shown on your first visit. Here's what that means in practice:

• Consent-gated by default. Google's Consent Mode v2 is wired in with all four consent signals (ad_storage, ad_user_data, ad_personalization, analytics_storage) defaulted to denied before any analytics call is made. Nothing is granted until you click "Accept" on the banner; clicking "Reject" (or simply not choosing) leaves everything denied. (See the Advanced Consent Mode caveat a couple of bullets down -- "denied" here doesn't mean literally nothing is ever sent to Google before you choose.)
• Why ad-related signals, on a site with no ads. Consent Mode v2 defines a fixed set of four signals as a package -- ad_storage, ad_user_data, ad_personalization, and analytics_storage -- and GA4's standard integration (the one wired in here) grants all four together when you click "Accept," even though this site only uses GA4 for basic analytics. That's the standard shape of Google's own consent-signal API, not a signal that this site runs ads or remarketing: there are still no advertising pixels or ad campaigns active on this site (see "No other tracking" below).
• Your choice is remembered. Once you accept or reject, that choice is stored in your browser (localStorage) so the banner doesn't reappear on later visits. If your browser blocks or clears that storage, the banner will simply show again.
• A real caveat about "denied": Consent Mode v2 here runs in Google's "Advanced" mode, not "Basic" mode. In Advanced mode, Google's gtag.js script loads for every visitor -- even before you make a choice, and even if you reject -- and Google may still receive limited, cookieless signals used for conversion modeling. This is different from analytics tracking being fully absent pre-consent; it's Google's own documented behavior for sites running Consent Mode v2 in Advanced mode rather than blocking the script entirely until consent is granted. We're documenting this plainly rather than implying "denied" means "nothing is ever sent."
• No other tracking. Google Analytics, as described above, is the only analytics or tracking mechanism on this site. There are no advertising pixels, session-replay tools, or third-party trackers beyond it.

We don't claim GDPR, CCPA, or any other privacy-law "compliance" as a certified or guaranteed status -- the above describes the site's actual, current behavior, not a legal conclusion about any specific jurisdiction's requirements.

For how Google itself describes its own data handling, see Google's Privacy Policy (external site, not something this project controls or vouches for). If you have privacy-related questions about this site specifically, please open an issue.

General liability

This site, its documentation content, and the kmp-user-agent library itself are provided on an "as is" basis, without warranty of any kind, express or implied, to the fullest extent permitted by law. See License for the library's own MIT license text, which spells this out formally for the code itself.

A specific consequence worth calling out for a User-Agent parsing library: a User-Agent string is client-supplied data, and the sender can set it to anything -- including a spoofed or forged value that doesn't match the actual browser, OS, or bot making the request. kmp-user-agent's parsed (or generated) output reflects only what the input string says, not any independently verified fact about the requester. It should not be relied on as the sole basis for security-critical decisions (for example, bot-blocking or fraud detection) without additional verification.

Docs content licensing

The library's own code is MIT-licensed (see License). Whether a separate license applies to this documentation site's own written content (the guide prose, page copy, and demo component source under docs-site/) is, as of this writing, an open question that hasn't been decided -- it isn't released under MIT, CC-BY, or any other specific license at this time. Don't assume any particular license applies to this content beyond what's stated here.