URL Extractor

A url extractor is a tool that scans any text input and pulls out every URL it contains. It uses a regular expression to find patterns starting with http, https, ftp, mailto, or www., then validates each match against the browser's URL parser to drop malformed strings. After validation, the tool deduplicates the list and groups results by host. This BlazeHive extractor handles 4,000-line log files, JSON payloads up to a few megabytes, raw HTML, and pasted prose without uploading anything to a server. Everything runs in the browser. Common use cases include security log triage, support ticket review, content migration audits, and quick competitive research on outbound links from a competitor article. Use the link-extractor when you specifically need anchor text and nofollow flags from a live web page instead.

Paste your text into the URL or paste text/HTML field, choose a filter, and click Extract URLs. The tool runs a regex pattern over your input that matches http://, https://, ftp://, mailto:, and bare www. strings. Each match runs through new URL() to filter out malformed candidates, then trailing punctuation like commas, periods, parentheses, and brackets gets stripped. The deduped list groups by host and shows a count for each domain. A 200-line email thread typically yields 5-15 unique URLs in under a second. For programmatic extraction, the same regex pattern works in Python (re.findall), JavaScript (String.prototype.matchAll), or grep with the -oE flag, but a browser tool gives you instant filtering and host grouping that a one-liner does not.

Yes. Server logs are one of the most common inputs for this url extractor. Paste up to a few megabytes of nginx, Apache, or application log lines into the textarea and click Extract URLs. The tool finds every URL referenced in the log, dedupes, and groups by host. A 4,000-line nginx access log with 12,000 raw matches typically reduces to 80-200 unique URLs after dedupe. For files larger than 5MB, split the log into chunks before pasting. If you need to extract URLs from logs as part of an automated workflow, the same regex pattern works in grep -oE on the command line. Use the browser tool when you want quick visual grouping. Use grep when you want to pipe the output into another script. After extraction, run suspicious external hosts through the link-extractor to inspect the actual page content.

The url extractor works on text, so first convert the PDF to text. On macOS, run pdftotext input.pdf output.txt from the command line (install via brew install poppler). On Windows or Linux, use the same pdftotext utility from the Poppler package. Open the resulting text file, copy the contents, and paste into the URL or paste text/HTML field. PDFs created from web pages typically contain 20-100 URLs. PDFs from native documents like reports usually contain 5-20. The extractor catches URLs that span line breaks if the PDF preserves them as continuous strings. Watch for PDFs that wrap long URLs across multiple lines with hyphens. Those become two broken URLs. If your output looks suspicious, open the PDF in Preview or Adobe Reader and use the built-in "Save as Text" export, which preserves URL continuity better than command-line tools for some PDFs.

Open the email, view the raw source or "Show Original" (Gmail) or "View Source" (Outlook), and copy the full HTML body. Paste that into the URL or paste text/HTML field and click Extract URLs. A typical marketing email contains 15-40 URLs including tracking pixels, unsubscribe links, social icons, and content links. Choose the HTTP/HTTPS only filter to skip the mailto: reply-to address. For Apple Mail or Thunderbird, you can copy the rendered email directly without viewing source, but you will miss the URLs hidden in image alt attributes and tracking pixels. View source for the most complete extraction. If you process emails in bulk, export your inbox to mbox format and run the file through the extractor in chunks. The same regex catches URLs in plain text and HTML email bodies equally well.

Every regex match runs through JavaScript's new URL() constructor. If the constructor throws, the candidate gets dropped. This filters out malformed strings like https:// with no host, http://example with no top-level domain, and partial fragments. The validator does not check whether the URL resolves or returns a 200 response. It only verifies syntactic validity. To check whether URLs are live, run the extracted list through the http-status-checker one host at a time. The validation step typically rejects 5-10% of raw regex matches as malformed. Bare www.example.com references get normalized with an https:// prefix before validation so they pass the parser. Trailing punctuation like )., ,, and ] gets stripped before validation to prevent false rejections from URLs glued to surrounding prose.

The extractor matches five protocol patterns: http://, https://, ftp://, mailto:, and bare www. (which gets normalized to https://www.). That covers 95% of URLs found in real-world inputs. The extractor does not match tel:, sms:, file://, chrome-extension://, or other browser-specific schemes by design. If you need to extract phone numbers from text that mixes URLs and tel: links, use the phone-number-extractor on the same input. The HTTP/HTTPS only filter restricts output to web URLs and drops mailto: and ftp:// matches. The All URLs filter shows everything. The 5-protocol scope keeps the regex pattern fast enough to process multi-megabyte inputs in under a second on modern browsers. Custom protocols can be added by editing the regex pattern in the source if you fork the tool.

Dedupe runs automatically after extraction. The tool compares each URL by full string match including protocol, host, path, query, and fragment. Two URLs that differ by even one character count as separate entries. That means https://example.com/page and https://example.com/page/ are two entries (trailing slash difference). It also means example.com/page?utm_source=email and example.com/page?utm_source=social are two entries. If you want unique pages regardless of query parameters, paste the output into a spreadsheet, split on ?, and dedupe the path-only column. A 12,000-match raw extraction from server logs typically dedupes to 80-200 unique URLs, a 60-150x reduction. The dedupe step is what makes the extractor useful for log triage. Without dedupe, you get the raw match count, which is rarely the number you want.

Paste the text, click Extract URLs, and the tool shows the unique URL count next to the host group totals. For raw match count before dedupe, look at the result summary which displays both numbers. A 1,000-word blog article typically contains 5-20 URLs. A 10-page legal document contains 30-80 URLs. A server log file contains 50-500 unique URLs depending on traffic patterns. If you need just the count without seeing the URLs, this tool still gives you the fastest answer. For programmatic counts in scripts, use grep -oE 'https?://[^ ]+' input.txt | sort -u | wc -l on the command line. The browser tool wins on inputs you already have in your clipboard. The command-line approach wins for batch processing 100+ files.

A url extractor reads raw text with regex and pulls out anything that looks like a URL. It works on logs, JSON, prose, broken HTML, and any unstructured input. A link extractor parses valid HTML with DOMParser and reads <a href> tags, returning structured metadata like anchor text, nofollow flags, and internal-versus-external classification. The url extractor is text-first. The link-extractor is HTML-aware. Use the url extractor when your input is messy or you do not have a live URL. Use the link extractor when you want to audit the link structure of a real web page with full anchor context. The two tools complement each other: extract URLs from a log, then run each unique URL through the link extractor to see what each page links to. Together they cover both unstructured text and structured HTML inputs.

The url extractor does not classify links by default because it has no concept of a base URL when you paste raw text. The Hide same-domain filter does the next-best thing: when you fetch a URL using the field above, the tool knows the source host and can hide URLs pointing to that same host. That gives you an external-only view. For pasted text without a fetched source, every URL counts as external because there is no base URL to compare against. If you need true internal-versus-external classification with anchor text, use the link-extractor. It fetches the URL, parses the HTML, and labels each <a> tag as internal or external based on the source page's host. A typical B2B blog post has 80% internal links and 20% external. A news article reverses that ratio.

Paste the raw HTML source into the URL or paste text/HTML field and click Extract URLs. The regex catches URLs inside <a href>, <img src>, <link href>, <script src>, <iframe src>, inline CSS url() calls, and plain-text URLs in the HTML body. A 50KB HTML page typically contains 80-200 unique URLs across all those sources. Use the HTTP/HTTPS only filter to focus on web URLs and skip mailto: and ftp:// matches. For HTML where you specifically want only the anchor links and need anchor text, use the link-extractor instead. It parses the DOM and returns only <a href> matches with text content. The url extractor is broader. The link extractor is more precise for anchor-specific audits. Pick based on whether you want every URL or only the anchor links.

The pattern matches the five supported protocol prefixes followed by valid URL characters: roughly (?:https?|ftp|mailto):\/\/[^\s<>"]+ plus a separate match for bare www. strings. The exact pattern handles edge cases like URLs ending in punctuation, URLs with parentheses, and URLs containing query parameters with special characters. After matching, each candidate runs through new URL() validation, so the regex can be slightly permissive without polluting the output. If you want to use the same pattern outside the browser, the simplified version https?:\/\/[^\s<>"]+ works in Python re.findall and grep -oE. The full pattern with mailto and www support requires more careful escaping. For most ad-hoc extractions, the simplified pattern catches 95% of real URLs. The browser tool uses the full version for completeness.

Yes. This url extractor is free with no signup, no upload, and no rate limit. Everything runs in your browser. You can paste up to a few megabytes of text and extract URLs in under a second. The tool does not send your input to a server, which matters when processing logs or emails that contain sensitive data. Compare to paid tools like Screaming Frog (250 free URLs, then $259/year for unlimited) or Ahrefs (no free URL extraction, requires $99/month subscription). Those tools include features the BlazeHive extractor does not, like recursive crawling and SEO scoring. For pure URL extraction from text or HTML, the free browser tool covers the use case. Pair it with the email-extractor and phone-number-extractor for full contact-data extraction from the same input, all free.

Because URLs in prose often have punctuation glued to the end. A sentence like "Read the article at https://example.com/post." has the period attached to the URL after a naive regex match. Without stripping, the extracted URL becomes https://example.com/post. which fails the URL parser because the trailing period is not part of the path. The extractor strips )., ,, ;, ], >, ", ', and . from the end of each match before validation. That recovers the clean URL https://example.com/post. The same logic handles URLs in markdown like [link](https://example.com) where the closing parenthesis is not part of the URL. Without trailing punctuation stripping, 10-20% of URLs extracted from prose would fail validation and get dropped. The strip step is what makes the extractor reliable on real-world text instead of just clean log files.