UserStack User-Agent Lookup MCP. Identify OS, browser, and device from any raw string.

When your AI agent runs into messy web traffic logs, it doesn't know what to do with raw User-Agent strings. That’s where the UserStack User-Agent Lookup server steps in. It uses a single tool, detect_user_agent, to instantly turn that jumble of text into structured data your agent can actually use.

This isn't just reading a string; it’s parsing complex client requests to figure out exactly who's making the call—whether it’s a human browsing on an iPhone or some automated search crawler.

Your AI agent gets granular details about the requesting device. The detect_user_agent tool first classifies the entire platform, telling you immediately if the request came from a mobile phone, a desktop computer, or a tablet. It handles the subtle differences between these categories so your agent knows how to adjust its behavior accordingly.

It pulls out deep technical details about the client's software. You’ll get the specific operating system—be it macOS, Windows, Android, or iOS—along with the full browser version number used by the client. This is crucial; knowing if a user is on Chrome 120 versus Chrome 123 changes how your agent needs to format outgoing requests.

Beyond just OS and browser versions, the server identifies the manufacturer's name. It reads the string to provide the brand data for the device, giving you names like Apple or Samsung. This level of detail helps your agent profile user hardware reliably.

The tool also checks if a request is coming from a known automated source. It detects whether the call originates from a human user or an established search engine bot, such as Googlebot. Your agent gets clear binary data on this front: human interaction versus programmatic scraping. This capability helps you differentiate between real-world usage and algorithmic polling.

The entire process revolves around taking that messy User-Agent string—the raw input—and converting it into clean, usable parameters for your agent to consume. The detect_user_agent tool doesn't just guess; it systematically structures the data points. It takes all the disparate pieces of information—platform type, OS version, browser name, device brand, and bot status—and packages them together in a format your AI client can process instantly.

If you’re auditing web traffic patterns or building personalized user experiences, this is what happens: Your agent sends the raw User-Agent string to detect_user_agent. The tool runs its parsing logic, returning structured fields. You immediately know if the request was initiated by a desktop Chrome user running Windows 11, or maybe an Apple Safari browser on an iPadOS device.

It doesn't leave anything vague; it gives you specific versions and brands.

This detailed breakdown lets your agent act like a dedicated network analyst when reviewing logs. Instead of just seeing 'browser detected,' the agent sees 'Chrome vX.Y running on Windows 10.' This specificity allows developers to write highly targeted code paths or adjust data models based on precise client environments. It's robust enough that if you feed it an unusual combination—say, a niche embedded device browser—it still provides the best classification possible across all known metrics.

It’s essential for debugging where platform-specific bugs are hiding, and for building APIs that need to behave differently depending on whether they detect Googlebot or a user with specific mobile hardware. The detect_user_agent tool handles this complex differentiation instantly, ensuring your agent's actions are always informed by the precise client context.