HAR Diff Analyzer

HAR (HTTP Archive) is a JSON-format file that records every HTTP request and response made during a browser session, including headers, timing breakdowns, status codes, and body sizes. To export one from Chrome or Edge: open DevTools (F12), go to the Network tab, reproduce the flow you want to capture, then right-click any request and choose 'Save all as HAR with content'. In Firefox, click the gear icon in the Network panel and select 'Save All as HAR'.

Requests are matched by HTTP method + URL path + query string, ignoring the host and protocol. This means a GET /api/users?page=2 in File A will be matched to the same path in File B even if the domain changed. If a URL appears multiple times in one file (e.g. polling), the first occurrence is used for comparison.

Red delta values mean the request got slower in File B compared to File A. Green delta values mean it got faster. Rows with a yellow background indicate a status code change (e.g. 200 → 404). The regression table is sorted by absolute timing delta, so the most impactful changes appear at the top.

No — all parsing and diffing runs entirely in your browser using JavaScript. Your HAR files never leave your machine. This is important because HAR files often contain sensitive data such as authentication tokens, cookies, and API keys.

The 'Changed' tab filters to requests with a timing delta of ≥ 50 ms or a status code change, which is a practical starting point. For user-facing page loads, regressions above 100 ms are typically noticeable. For API calls on a critical path, even 20–30 ms differences can compound significantly under load. Use the 'Performance Regression' tab to see the full sorted list and decide your own threshold.