Parse a user-agent string online when debugging browser analytics, testing server-side classification, or trying to understand a client-identification header. This tool reports approximate browser name and version, operating system and version, device category, and rendering engine based on a small set of regular-expression rules. It is aimed at web developers, QA teams, and learners investigating compatibility issues. A user-agent string is self-reported text and can be shortened, spoofed, or changed, so the displayed labels are clues rather than proof of a visitor's identity or capabilities.
The field labelled “User Agent String” is prefilled from the current browser's navigator.userAgent when available. You can keep that sample or paste another string, then choose “Parse User Agent.” The result panel displays parsed fields; an empty field disables the action. For a test string containing Chrome/120.0 and Windows NT 10.0, the rules identify Chrome, Windows with the parser's 10/11 label, a desktop device, and a WebKit engine when AppleWebKit is present. Clear removes the current text and result so you can test another sample.
The utility checks selected markers in a fixed order: Edge, Opera, Chrome, Firefox, Safari, and Internet Explorer patterns for browser detection; several Windows, macOS, Android, iOS, Linux, and Chrome OS patterns for platform; simple Mobile and Tablet indicators for device; and Gecko, AppleWebKit, or Trident markers for engine. That order matters because Chromium-based browser strings often contain Chrome tokens too. The tool exposes no current browser database, client-hint parser, custom rules, confidence score, or detailed hardware inventory.
Use the output to make sense of a log line or compare representative test strings, not to make high-stakes access decisions. Browser vendors intentionally include compatibility tokens, versions can change rapidly, and privacy tools may reduce or modify the string. The parser may classify Windows NT 10 as Windows 10/11, which explicitly combines two releases, and it defaults unknown device values to Desktop. A result of Unknown means the rules did not recognize a pattern, not that the source has no browser or operating system.
When checking a production issue, compare multiple known samples and test the actual feature rather than relying on a parsed label. A server may receive a different header than the browser's current navigator.userAgent if a proxy or client has altered it. The HTTP Header Viewer provides static explanations and sample request headers, not a capture of your current network request. For network-origin questions, the IP Lookup page separately asks an external service for the public IP; user-agent parsing itself does not need that lookup.
The component initializes the field from navigator.userAgent when rendered in a browser and passes the entered string to the local parser; no network request for parsing is shown. This is specific to the component and not a wider privacy guarantee. A user-agent string can itself reveal software details, so avoid sharing unredacted logs unnecessarily. Use the tool for diagnostics, document its rule limitations, and verify compatibility with feature detection and real browser tests where possible.
The parser reads the string supplied in the text area, so it can also help compare a captured server log with the current browser's own value. Those sources may differ because a proxy, application, privacy setting, or test harness changed the header. Record where a sample came from when investigating a report, and do not assume the auto-filled field reproduces the exact header seen by a remote service.
Several browser names share engine tokens, and the helper uses recognition order to make a best-effort label. It does not report installed extensions, hardware model, exact screen size, or every browser release. Device classification relies on a few string patterns and defaults to Desktop, so tablets and unusual embedded browsers may not be recognized consistently. A blank version does not mean an application is unversioned; it means no supported version pattern matched.
For compatibility analysis, build a small set of real test strings from browsers your product supports and compare the results with actual feature tests. Avoid using user-agent parsing to fingerprint people or to deny functionality based on weak heuristics. Feature detection, progressive enhancement, and standards-based behavior usually provide a more reliable path than classifying an opaque string. The parser is best used to understand logs, not to make decisions about an individual's trustworthiness.