Generate an XML sitemap from your addresses

How to use the tool

Paste one address per line, or open a text or CSV file exported from your CMS or a crawler. If you know when a page was last modified, put the date in a second column, separated by a semicolon or a tab. A click on "Generate the sitemap" shows the findings, then the files to copy or download, and the line to add to robots.txt.

Enter the folder where you will upload the files if it is not the site root: it is used to write the index addresses and the robots.txt line. The "Load an example" button fills in a list with a few common mistakes and sets four addresses per file, to show the splitting and the index.

What the tool cleans up

The protocol requires a UTF-8 encoded file and escaped values: &, ’, ", > and < become &amp;, &apos;, &quot;, &gt; and &lt;.¹ Google gives the same rule for all XML files.² An address with a parameter such as ?colour=red&size=42 would otherwise be unreadable, and Search Console names an unescaped character as a common cause of parsing errors.⁴

Addresses must also be encoded according to the URI standard: the page ümlat.php becomes %C3%BCmlat.php in UTF-8.¹ The tool encodes accented characters, spaces and disallowed signs, lowercases the host, writes an accented domain name in its ASCII form and shows two rewriting examples. An address must be under 2,048 characters: beyond that, it is removed.¹

A fragment (#team) points to part of a page: the tool removes it, then deletes exact duplicates and keeps the first line. Relative addresses, such as /contact/, are dropped: Google asks for fully qualified, absolute addresses, which it crawls exactly as listed.²

A single host, a single protocol

All the addresses in a sitemap come from the same host, www.example.com or store.example.com, and use the same protocol.¹ The tool takes the host of the folder you entered, otherwise the most frequent one in the list, and removes the others by name: example.com without www, a subdomain, or a page still on http. Make one file per host.

Unless it is submitted in Search Console, a sitemap only affects the descendants of its folder: Google therefore recommends the site root.² If you enter a folder such as /maps/, addresses located elsewhere are flagged.

lastmod dates

The date follows the W3C Datetime format: YYYY-MM-DD is enough, the time is optional.¹ If a time is given, so is the time zone: 2026-10-21T10:00 is rejected, 2026-10-21T10:00+02:00 is accepted.⁴⁵ The tool accepts the six forms of the W3C note, checks months, days and hours, and keeps the address without a date when its date is wrong. A date such as 03/14/2026 is flagged.

Google uses lastmod if it is accurate and verifiable, and it should reflect the last significant update of the page: main content, structured data or links; the copyright date does not count.² Bing may disregard dates that are always today and asks not to use the sitemap generation time.⁶⁷ The tool’s shared date therefore only applies to addresses without a date, and the finding reminds you that it is only right if those pages really changed that day. A date in the future is flagged.

Splitting and the index file

A sitemap holds at most 50,000 addresses and 50 MB uncompressed; beyond that, you need several files.² The protocol gives the exact limit: 52,428,800 bytes.¹ The tool counts both while writing and splits at the first limit reached. You can also choose fewer addresses per file: Google notes that several sitemaps let you track the search performance of each one in Search Console.²

The files are then listed in a sitemap index, with the same format and namespace, which can contain up to 50,000 loc tags.³ Listed files must be in the index’s folder or lower: the tool puts them all in the same folder.³ An index cannot list another index.⁴ The optional lastmod tag of the index refers to the modification of the sitemap file itself: the tool does not write it, since it does not know when you will upload the files.³

priority and changefreq

The protocol describes both tags as hints: changefreq is "considered a hint and not a command", and priority defaults to 0.5.¹ Google says it in one sentence: "Google ignores <priority> and <changefreq> values".² Bing wrote in 2025 that these tags are ignored and affect neither crawling nor ranking.⁷

The tool therefore does not write them by default. The unchecked option adds them with the same value on every address, for an engine or an internal tool that would still read them. Bing had found in 2023 that most of these tags were set incorrectly, often with the same value across the whole sitemap.⁶

Declaring the sitemap

Upload the files to the folder you entered, then add the suggested "Sitemap:" line anywhere in robots.txt: Google finds it on its next visit, and several lines are allowed.² If there is an index, declaring the index alone is enough, without listing each file.¹ You can also submit the address in the Sitemaps report of Search Console.²

Limits of the tool

The tool visits no page: it formats the list you give it. It cannot tell whether an address returns 200, redirects, carries a noindex or a canonical pointing elsewhere. Yet Google asks you to list the addresses you want to see in its results, the canonical version when content has several addresses.²

It writes uncompressed XML files, without the image, video, news or hreflang extensions. Two addresses that differ only by a trailing slash remain two distinct addresses. The CSV is read on two columns, the address then the date.

After generating

Once the files are online, run them through the XML sitemap validator: it reads robots.txt, the index and a sample of addresses. For the "Sitemap:" line and the crawl rules, the robots.txt generator writes the full file. And to see what Google really does with a sitemap, read the article on XML sitemaps in 2026.

Frequently asked questions