What This User-Agent Parser Does
Use this browser-based tool to parse one User-Agent string or several lines of input. It returns browser, rendering engine, operating system, device type, device model where detectable, bot status, the raw string, and some unknown tokens.
Enter User-Agent Strings
Paste a single User-Agent string below, or put one string on each line for bulk parsing. Empty lines are ignored. Results update automatically as you type; Add my browser's UA appends the current browser string.
Review Parsed Results
Each result row can show the browser and version, rendering engine and version, operating system and version, device type, device model, bot or crawler status, unknown-token count, and a shortened raw string. Bot detection runs first, so a bot row may not receive a browser classification.
Click the chevron button in the first column of a result row to open its details. The expanded view includes the full raw User-Agent string and some remaining Name/Version-shaped tokens that the parser did not consume. Detection is best-effort, so an unclassified token is not guaranteed to appear there.
Copy or Export Results
Use the visible controls to move the parsed rows elsewhere:
- Copy JSON, Copy JSONL, or Copy CSV copies the results.
- Export JSON, Export JSONL, or Export CSV downloads the results.
JSONL writes one JSON object per line.
Clear the Workspace
Click Clear to empty the input, remove the current results, and hide the workspace and result controls.
Detection Limits and Browser-Local Processing
The parser uses signature-based, best-effort detection rather than an exhaustive User-Agent grammar. It can identify common browser, engine, operating-system, device, and bot patterns, but unusual or deliberately obfuscated strings may remain unrecognized. Some remaining Name/Version-shaped tokens may be listed as unknown tokens; the tool does not guarantee that every unclassified token will be surfaced.
The parsing and export logic runs in the browser, and the local implementation does not use a remote lookup service for parsing. The tool reads the text you supply; it does not inspect live network requests, client-hints headers, or traffic, perform IP geolocation, or generate and spoof User-Agent strings.
Frequently Asked Questions
Does this tool detect browsers from live network requests?
No. It only parses the User-Agent string text you paste in - it does not inspect incoming HTTP requests, client hints headers, or any live traffic.
Does it do IP geolocation?
No. This tool has no concept of IP addresses at all - it decomposes the User-Agent string only. For geolocation you'd need a dedicated IP lookup tool.
Can it generate or fake a User-Agent string?
No, this is a one-way parser. It reads a User-Agent string and reports what it can determine about it - it does not construct or spoof new UA strings.
What happens to tokens it can't classify?
Some remaining Name/Version-shaped tokens may be listed as "unknown tokens" on that row. This is a best-effort check, not a guarantee that every unclassified token will appear.
How are bots and crawlers identified?
Against a built-in signature list covering common search engine crawlers, social-media link previewers, SEO tools, AI/LLM crawlers, and generic HTTP clients (curl, Wget, PostmanRuntime, headless browsers). A match short-circuits browser detection, since bot UAs often embed browser-looking tokens for compatibility.
Can I parse many User-Agent strings at once?
Yes. Paste one User-Agent string per line for bulk parsing - each line becomes its own row in the results table, and exports include every parsed row.