A HAR network recording
A standard file exported from your browser after you reproduce a broken checkout, login, upload, dashboard, or other web flow.
Free browser-error analyzer · 0 bytes uploaded
A website broke.
Send evidence, not guesswork.
A HAR is the browser's recording of what happened while a page failed. ReqRescue finds the likely failure points, removes common secrets, and turns that recording into a ready-to-send bug report. The file never leaves this tab.
0bytes uploaded
< 10sto first diagnosis
3export formats
Drop the browser recording (HAR).
Nothing is uploaded. Your browser performs the entire analysis and redaction locally.
Chrome, Firefox, Edge, Safari · protected local limit 75 MB per file · for A/B, choose baseline first and changed capture second
Need a HAR? Export one in 3 steps.
A HAR is a browser recording of the network requests made while a page loads or a problem happens. It helps support and engineering see failed API calls, slow requests, redirects, and status codes.
- Open the broken page, then open Developer Tools.
- Select Network, reload the page, and reproduce the issue.
- Export or save all requests as HAR, then open that file here.
00
ReqRescue, in plain English
A safer, faster handoff when a website breaks.
ReqRescue reads a browser network recording, finds the requests most likely connected to the failure, removes common secrets, and packages the useful evidence for the person who has to fix it.
Support, QA, developers, and technical founders
Anyone who needs to turn “it does not work” into evidence an engineer can inspect without another round of questions.
A diagnosis, clean HAR, and ready-to-send report
Ranked suspects with confidence, cited request evidence, detected privacy risks, and a Markdown handoff for Jira, GitHub, Slack, or AI.
01
The support ritual, compressed
Sanitizing is step one.
Explaining is step two.
A viewer hands you hundreds of rows. A sanitizer removes risk. ReqRescue does both, then performs the first triage pass an experienced support engineer would do—and shows its work.
Reconstruct the failure
Group repeated failures, trace status patterns, isolate slow-tail requests, and spot redirect loops that keep the browser in circles.
Separate fact from hypothesis
Every suspect carries a confidence level and the exact requests that support it. No invented root cause.
Make the evidence shareable
Strip cookies, auth headers, tokens, passwords, emails, IPs, private hosts, and sensitive body fields.
02
What changes
POST /v1/checkout 401 462ms
Authorization: Bearer eyJhbGc…
Cookie: session=sess_live_9f8…
POST /v1/auth/refresh 401 390ms
Set-Cookie: refresh=rf_2ba…
… 184 more requestsRisky to share. Slow to understand.
Likely auth refresh failureHigh confidence
- 3 matching 401 responses
- First failure: POST /v1/checkout
- Session + 12 secrets redacted
Ready for Jira, GitHub, Slack, or an AI debugger.
Privacy architecture
Your trace never takes a network trip.
Parsing, triage, and file generation run in this tab. We record only pseudonymous product events such as “demo analyzed” or “report exported,” using a random identifier stored in this browser—never file names, URLs, headers, bodies, or HAR contents.
Your HAR→Your browser→Your exports
03
Trust, before upload
The questions a sensitive
trace deserves.
ReqRescue is deliberately local, deterministic, and explicit about uncertainty. No account is required.
Does ReqRescue upload my HAR file?
No. Parsing, ranking, redaction, and export generation happen in your browser tab. The server never receives the HAR, its file name, request URLs, headers, or bodies.
What sensitive data does it remove?
ReqRescue strips request and response bodies, cookies, authorization headers, API keys, access and refresh tokens, passwords, emails, IP addresses, private hosts, URL credentials and likely identifiers in paths. It also drops unknown vendor fields. Automated redaction cannot guarantee recognition of every product-specific secret, so review every export before sharing it.
How is this different from a HAR viewer or sanitizer?
A viewer helps you inspect requests. A sanitizer removes secrets. ReqRescue also groups failures, ranks evidence-backed suspects, and produces a Markdown incident report plus a clean HAR for engineering or support.
Does it claim to know the root cause?
No. ReqRescue separates observed facts from hypotheses. Each suspect has a confidence level and the exact requests that support it, so an engineer can verify the diagnosis instead of trusting a black box.
Can I try it without a real trace?
Yes. Run the 15-second broken-checkout demo above. It uses a synthetic HAR with repeated authentication failures and fake secrets, so you can inspect the complete report safely.
ReqRescue is free · pass the help forward
Saved an hour? Help a cat instead.
ReqRescue does not accept donations. Analysis, comparison, redaction, PDF, local history, and exports are free. If the tool saved you time, consider donating directly to one of these independent Israeli animal organizations.
These links go straight to the organizations' official donation pages. ReqRescue is not affiliated with them and receives nothing.No paywall, license key, account, subscription, or donation to us.
- Analyze or compare local HAR files up to 75 MB each
- Download clean HAR and Markdown reports
- Print or save full incident briefs as PDF
- Save 10 case reports locally on this device