ua-tracer is a tool designed to help web developers understand the specific actions taken by different user agents—such as browsers, bots, indexers, and scrapers—when they access a web page. It addresses the challenge of determining whether an agent merely fetches the HTML, parses additional resources like CSS or fonts, follows links within those resources, or actually executes JavaScript. The tool provides detailed, request-by-request traces that reveal which assets an agent fetched, which resources it followed, and whether it performed actions such as running scripts or posting client-side performance data.
When a user agent visits the associated site, ua-tracer generates a unique trace ID and a per-request secret, embedding these in the paths of every asset (including stylesheets, scripts, images, fonts, and manifests) served for that session. This mechanism ensures that only the original agent can fetch the probe assets, preventing replay or guessing by others. The tool sets up additional probes within assets to detect whether the agent parses CSS and follows background images or font sources, parses a manifest and follows icon references, or executes JavaScript (both classic scripts and ES modules). It also detects if the agent posts back client-side resource timing data, indicating a real browser engine ran the page.
The trace detail for each session is publicly accessible via a shareable link, making it easy to communicate findings, such as whether a particular agent executed JavaScript or only fetched static resources. ua-tracer also checks the source IP of each request against published CIDR ranges from major crawler operators, helping to verify whether a trace genuinely originates from a known bot or is likely spoofed based on the User-Agent string and IP match. The tool has been used to observe and compare behaviors of various bots, including those from OpenAI and Anthropic, revealing differences in how deeply they interact with web content.
ua-tracer is particularly useful for web developers and researchers seeking insight into how different automated agents and browsers interact with their sites. By providing granular, session-specific traces and agent verification, it supports analysis and debugging of web resource access patterns.
ua-tracer is a Developer Tools product. It focuses on understanding what resources browsers, bots, and AI agents actually fetch and execute when loading a web page. It is built as a B2B product for web developers. The product is available for free. It runs on the web.
Behind ua-tracer is Paul Kinlan, and the product first shipped in 2026. Development happens publicly on GitHub with 26 commits in the last 90 days. PulseGate's similarity index finds few close equivalents — ua-tracer occupies a relatively distinct niche. Key capabilities include request tracing, user agent analysis, and resource fetch log.
Latest indexed changes and source events
What does your user agent do? discovered by the PulseGate indexer
Other apps tracked under the same category.