HAR File Viewer and Analyzer
Inspect HAR requests, separate body/content sizes and supported cURL snippets
A HAR may contain cookies, authentication headers, URL parameters and bodies. It is read in your browser without upload or draft storage; copied cURL may also contain these values.
About this tool
HAR File Viewer reads an HTTP Archive locally and displays searchable requests, status codes, headers, recorded response body sizes, decoded content sizes and durations. Unknown values stay unknown instead of being substituted with a different metric. The archive is limited to 30 MiB and 10,000 requests. Selected entries can produce a POSIX cURL snippet when their method, URL, headers and text or URL-encoded body can be represented. Multipart files, missing bodies, invalid request fields and oversized snippets are explained without pretending to reconstruct the full request. This is an archive inspector, not a replay client or a complete HAR schema validator. Files are not uploaded and no archive drafts are stored.
Production Snippets
Unknown wire body size is not the decoded content size
text
{
"response": {
"bodySize": -1,
"content": { "size": 2400, "mimeType": "application/json" }
}
}
// Recorded response body size: unknown
// Decoded content size: 2400 bytes
// Do not report 2400 bytes as measured network transfer.Frequently Asked Questions
Which files can be opened?
JSON HAR archives with log.entries and request/response records are supported, up to 30 MiB and 10,000 entries. Required fields and displayed numeric values are checked; this is not a full HAR schema validator.
Why are body size and content size different?
HAR response.bodySize records the response body size, while content.size can describe decoded content. Compression may make them differ. Unknown -1 values remain unknown; body size excludes response headers and is not total network transfer.
Does the duration sum equal page load time?
No. Known request durations are added and unknown durations are counted separately. Requests can overlap, so their sum does not measure wall-clock page load time.
Which bodies can become cURL?
Recorded text and URL-encoded text parameters are supported, including repeated values and empty text. Missing bodies, multipart files and unsupported encoded bodies do not produce a complete snippet.
Will the snippet replay the archive exactly?
No. It preserves supported method/header/body fields, but cURL manages Host, length, connection, encoding and HTTP/2 pseudo-headers. Authentication state, sessions, timing, transport defaults and server responses are not reproduced.
Does the tool upload or save the HAR?
No. The file is read only in the browser and is not saved as a draft. It can contain cookies, tokens, URLs and private bodies; copied cURL can contain them too. Clearing or replacing a file invalidates a pending read.
Keep browsing