Free URL, Link and Domain Extractor

The input mode determines what the extractor looks for. Select the format that matches your source rather than expecting every type of reference to be treated as a website link.

In Text mode, an email address or an unprefixed domain such as example.org is not a URL match. Use List mode for one bare hostname per line, or choose Mixed mode deliberately to review supported bare-host and email-derived domain evidence.

Local files must contain supported UTF-8 text, HTML, Markdown, CSV, JSON or XML. CSV text is not an Excel workbook decoder. The tool does not decode PDFs, Word documents, Excel files, email MIME attachments or ZIP archives.

1. Supply your source. Paste a supported format or choose a local UTF-8 source file. Keep meaningful source content and inspect any format error before extracting.
2. Select the matching mode. Use Auto detection as a starting point, then confirm its choice. HTML, Text and List keep their own rules; Markdown, CSV, JSON, XML and Mixed modes have the boundaries described above.
3. Set a base when needed. For relative HTML references, enter the actual HTTP(S) page URL that provides the intended context.
4. Review the result and diagnostics. Check assumptions, rejected candidates, counts, unresolved references and whether the report is complete or partial.
5. Copy or download the appropriate scope. Keep the full-report JSON or full-view CSV when you need every retained result. For a smaller handoff, choose the filtered or selected-row scope and review its row count before copying or downloading.

Leave query and fragment removal off when those parts may matter. Review the effect of an optional cleanup before using the resulting list elsewhere.

Consider this input in Text mode:

Visit https://example.com/Guide?ref=one#intro, and www.example.org/help.
Again: https://example.com/Guide?ref=one#intro.
A bare example.net and email [email protected] are not text-mode links.

With the default HTTPS assumption and cleanup disabled, the extraction engine returns two unique URLs from three accepted occurrences:

The sentence punctuation is excluded from the matched addresses. The bare domain and email are not extracted in this mode. This example was checked against the current downloaded extraction engine; it does not involve visiting either destination.

An HTML link such as href="/about/" does not contain its own hostname. Supply an explicit base to resolve it meaningfully.

With the base https://example.com/guides/draft/, standard URL resolution produces:

The default relationship check compares exact hostnames, so a different subdomain is external in this example. Optional www-alias or registrable-domain policies change that classification deliberately; read the selected policy in the report. Without a supplied base, relationships are unknown and relative references are unresolved. Embedded HTML base tags are ignored rather than silently overriding your chosen context.

This table illustrates the resolution and classification rules. It is not a response-status check.

Duplicate removal and URL cleanup answer different questions. By default, the extractor preserves differences that may identify distinct resources.

For example, these remain three different normalized URLs:

https://example.com/page?x=1#one
https://example.com/page?x=2#two
https://example.com/page

Choosing both remove query strings and remove fragments combines them into one URL with three occurrences. The report records the removals. That can be useful for a broad page-path review, but the original parameters or fragments may be important for products, attribution or document navigation.

For selective cleanup, choose the tracking-parameter keys you want removed instead of deleting the whole query. Inspect the original-to-output values and recorded changes; preserve functional keys such as product identifiers. Optional scheme or trailing-slash formatting is a rewrite, not proof of working HTTPS or the website’s canonical URL.

Review the parsed scheme, port, path, query and fragment alongside the normalized address. Query entries keep repeated keys and empty values. These fields explain the supplied URL’s structure; they are not an HTTP response or a canonical check.

The Public Suffix List contains rules for suffix boundaries. It allows the tool to handle structures such as .co.uk and private hosting suffixes more accurately than taking the last two labels of every hostname.

Private suffix rules are enabled by default. You can choose an ICANN-only grouping policy when that matches your task; hosted-tenant groups may then combine differently. Optional www removal is an output-formatting choice, separate from relationship classification. The report records the policy and bundled library version. Its suffix data is a pinned snapshot, not a live lookup.

IP addresses, localhost and unknown-suffix hosts may have no registrable-domain result. Internationalized hostnames retain normalized ASCII/punycode values, with Unicode display fields alongside them for review. Choose full host, registrable domain, subdomain, public suffix or final-label output; the final label is not the same as a registrable domain. A missing root or a recognized suffix is a classification result, not a DNS lookup, registration check or availability decision.

The Public Suffix List’s explanation describes this distinction and why the list should not be used as proof that a domain exists.

Use the HTML-link view to review the actual labels and attributes supplied in your source. Check whether a label explains its destination, whether a URL appears with several different labels, and whether a supplied rel value includes tokens such as nofollow, sponsored or ugc.

These attributes describe how the link is qualified. They do not provide a guaranteed measure of ranking value, and an absent nofollow token does not promise crawling or ranking benefit. Google’s guidance on qualifying outbound links explains the intended use of these values.

For more context around the wording of a link, use the Anchor Text Optimizer. Review a public page’s outbound-link counts and supplied rel information with the Outbound Link Counter.

Focus on internal or external relationships, supplied rel tokens, anchor wording or empty labels. Hostname filters distinguish exact matches from a deliberate domain-suffix match; a substring in an unrelated hostname is not enough. Public-suffix filters use the report’s suffix field, not a guess from the final two labels.

Sort by source order, address, hostname or frequency with stable ordering for ties. Counts distinguish retained occurrences, unique URLs and grouped hosts or domains. Changing a display filter does not make a partial scan complete.

Use the full export actions for every retained row, or choose an explicit filtered or selected-row scope for a smaller list. Review the scope, row count and columns first. Copy has a selectable-text fallback when clipboard access is denied. TXT, CSV and JSON serve common handoffs; supported TSV, Markdown, HTML, bookmark and JavaScript-array outputs apply their own escaping. A generated HTML file does not run inside the tool, and no native Excel workbook is claimed.

Your supplied text or file content is analyzed inside the isolated extractor workspace. That workspace is configured to block network connections during analysis, and supplied HTML is parsed without inserting it as live page content. Extracted values appear as text rather than links that automatically open.

The surrounding EduGuestPost website still makes normal website requests. The local-processing description applies to the tool’s handling of your supplied source, not to every request made by the parent page. See the site’s privacy policy for broader website information.

Reports do not include your original source document or automatically include local filenames. Generic source labels identify batch inputs unless you deliberately choose a label. Extracted URLs, raw references, labels, attributes and source metadata can still contain private information. A scoped export is not an anonymized export; review it before sharing.

The default workspace limits are 512 KiB of UTF-8 input and 500 processed candidates per run, shared by up to five supplied sources. Candidates include duplicates and rejected entries, so this is not a promise of 500 accepted URLs. Additional URL-length, HTML-node, time and output limits can also end a scan early. Check the active configured limits in the workspace.

A partial report contains the retained results and the reason processing stopped. It does not establish that the remaining source has no more URLs. Split a larger source into manageable parts if needed, and account for duplicates when reviewing several reports.

Results display in pages of 25 rows. The existing full-view CSV action exports every retained row in its view; full-report JSON retains the complete report. Separate scoped actions can export filtered matches or selected rows, with an explicit row count. Pagination alone does not limit an export. The default downloadable output limit is 2 MiB; the active workspace limit is shown.

The extractor does not fetch websites, crawl a site, test HTTP status, resolve redirects, check domain ownership, identify backlinks pointing to your site or confirm Google indexing. Its counts describe the supplied source within the report’s stated scope.