A HAR (HTTP Archive) file is a JSON document that logs the HTTP requests and responses a browser made while loading a page. Browser developer tools can export one, which makes it a common way to capture the exact requests a site sends so a scraper can reproduce them.
The widely used HAR 1.2 layout has a root object named log with a version, a creator, optional pages, and an entries array. Each entry describes one request and its response: method, URL, headers, cookies, query string, post data, status, content type, body, and timing. The file is UTF-8 JSON, so any language can read it with a normal JSON parser.
The format came out of a 2012 W3C Web Performance Working Group draft that never became a formal Recommendation, yet HAR 1.2 is what browsers and proxy tools export in practice. Treat the specification as a stable convention rather than a ratified standard, and expect some tools to add their own underscore-prefixed fields.
Exporting a HAR after a manual visit shows every request behind the page, including the background calls that fetch the data. Filtering entries by response content type is a fast way to find the JSON endpoint worth calling directly, and the recorded headers and cookies show what a working request looked like to compare against a failing script.
A HAR is evidence of one session, not a recipe. Tokens, timestamps, and signed parameters in it expire, and a replay from a different network or client can still be blocked. A HAR also stores cookies, authorization headers, and request bodies in plain text, so scrub it before sharing and never commit one containing a live session.
import json
with open("session.har", encoding="utf-8") as f:
har = json.load(f)
for entry in har["log"]["entries"]:
mime = entry["response"]["content"].get("mimeType", "")
if "json" in mime:
req = entry["request"]
print(req["method"], req["url"], entry["response"]["status"])How Crawlora handles this
A HAR shows you how one site's front end gets its data, but reproducing those calls means maintaining headers, sessions, and anti-bot handling yourself. For platforms Crawlora covers, its endpoints return structured JSON directly, so you can skip the capture-and-replay step; check the platform catalog to see whether your target is covered.
Related reading
FAQ
Open the browser's developer tools, load the page with the Network panel recording, then use the export option to save the log as HAR. Some browsers offer a variant that includes response bodies, which makes the file much larger.
No. A HAR is a JSON log of one browser session, mainly for debugging and inspecting requests. A WARC is an archival container designed to preserve captured responses at crawl scale.
Not without cleaning it. HAR files can contain session cookies, authorization headers, and form data. Remove or redact those values before sending one to anyone.
Beyond HAR File (HTTP Archive), Crawlora's own docs cover the rest of the stack — browse the APIs, test a request in Playground, and move from scraping infrastructure work to production data workflows.