In Google Search Console, “Sitemap could not be read” means Google tried to download your sitemap and failed, so it never got as far as the XML inside. On the main Sitemaps report the same failure reads “Couldn’t fetch”, and the fix nearly always sits on the request side: the address you submitted, the status code it returns, robots.txt or a firewall.
So start with the request, not the file. Each check takes a minute in a browser or a terminal, and Google’s help page for the report names five causes.
We build ContextBolt SEO, which lets an AI agent like Claude read your Search Console and audit your site, so weigh the bias. Everything below works with a terminal and Google’s free tools.
- What it means: Google could not download the file at the exact address you submitted. “Couldn’t fetch” on the list and “Sitemap could not be read” on the details page are one failure.
- Check first: request that address without following redirects. You want HTTP 200, an XML content type, and no robots.txt rule on the path.
- Google’s five causes: a robots.txt rule, a manual action, a wrong address (404), another server or HTTP error, and low crawl demand.
- A new sitemap: if the file loads and the URL Inspection live test passes, Google retries for a few days on its own. Resubmit once, then wait.
- Then check the pages: ContextBolt SEO runs a site audit from your AI agent that finds the sitemap the way a crawler does and flags listed pages that error or redirect.
What “sitemap could not be read” means
The Sitemaps report shows one status per sitemap, for Google’s most recent request. Google’s Sitemaps report help page, read on November 16, 2026, defines three.
| Status on the list | Details page says | What happened |
|---|---|---|
| Success | Nothing to fix | Google fetched and read the file with no errors |
| Couldn’t fetch | Sitemap could not be read | Google could not download the file |
| Has errors | Sitemap can be read, but has errors | Google downloaded it and some entries failed |
The split decides the fix. A broken XML file is a “has errors” problem: Google has the file and tells you which line it choked on. “Could not be read” means Google never had the file, so rewriting the XML changes nothing until the request works.
The details page also names the reason. Click the sitemap row and expand the section under the error. “HTTP error: 404” and “General HTTP error” each point to a row in the first table further down.
Check the request in four steps
- Copy the exact address from the report. Click the sitemap row and copy the URL Google shows. Google’s help says this column holds “the exact URL specified when the sitemap was submitted. Redirects are not followed.” An address you retype from memory proves nothing.
- Request it without following redirects. In a terminal,
curl -sIprints the headers of the first answer and stops there. You want200and a content type ofapplication/xmlortext/xml. A301or302means the submitted address is an old one: submit the address in theLocationheader instead. - Open robots.txt on the same host. Look for a
Disallowline that matches the sitemap’s path. In Google’s words, “Google respects robots.txt when fetching sitemaps.” ADisallow: /left over from a staging site blocks the sitemap along with everything else. - Run the live test in Search Console. Paste the sitemap address into the URL Inspection bar, click Test live URL and open Page availability. Google’s help says “Crawl allowed?” should read Yes and “Page fetch” should read Successful. Of the four steps, this one runs from Google’s own crawler.
Steps 2 and 3 as one paste, plus a parse check for later. Swap in your address and run it.
SITEMAP="https://example.com/sitemap.xml" # the exact URL from the Sitemaps report
# Status and type, without following redirects
curl -sI "$SITEMAP" | grep -iE '^(HTTP|content-type|location)'
# The first bytes: an XML declaration, then <urlset or <sitemapindex
curl -s "$SITEMAP" | head -c 120; echo
# robots.txt on the same host
curl -s "$(echo "$SITEMAP" | cut -d/ -f1-3)/robots.txt"
# Does it parse? Prints "parses", or the line and column that broke
curl -s "$SITEMAP" | python3 -c "import sys, xml.dom.minidom as m; m.parseString(sys.stdin.buffer.read()); print('parses')"
A curl request from your laptop is not Googlebot. A firewall or bot filter that blocks Google’s crawler can still answer you with a 200. Step 4 sees what Google sees. If curl passes and the live test fails, read why Googlebot gets a 403 while your browser loads fine: the rule is usually at your CDN.
To run these checks without a terminal, paste the address, or just your domain, into our free sitemap checker. It follows each redirect, reads robots.txt and the XML, and names the Search Console message each problem would cause.
We ran the check on our own sitemap
contextbolt.com is an Astro site served by Cloudflare. On November 16, 2026 we ran the script above on the address in our Sitemaps report, then requested the three addresses a person could plausibly submit instead.
| Address requested | What came back | Submit it? |
|---|---|---|
| contextbolt.com/sitemap-index.xml | 200, application/xml, a sitemap index listing one child file | Yes, and it is the one we submitted |
| contextbolt.com/sitemap-0.xml | 200, application/xml, 78,966 bytes, 695 URLs, all on contextbolt.com | Covered by the index |
| contextbolt.com/sitemap.xml | 301 to /sitemap-index.xml | No, it redirects |
| http://contextbolt.com/sitemap-index.xml | 301 to the https address | No, it redirects |
| www.contextbolt.com/sitemap-index.xml | No answer: the www host has no DNS record | No, and it belongs to another property |
The parse check printed “parses”, and robots.txt allows everything and names the index on a Sitemap: line. The Search Console API row for the submitted index shows 0 errors, 0 warnings and 695 URLs, last downloaded on November 9, 2026. It was last submitted on April 7, so Google keeps re-reading it on its own schedule with no resubmit. URL Inspection on the same address, read through the API, shows the last crawl on November 11: crawl allowed, page fetch successful, fetched by Google’s smartphone crawler.
So a working site still has three wrong addresses for its sitemap. /sitemap.xml is the one people guess, and on our site it redirects. Check where your platform writes the file before you type an address into the report.
Every sitemap error and its fix
These are the messages from Google’s Sitemaps report help page, read on November 16, 2026, split by what failed. The first table is “could not be read”: Google never got the file.
| Reason Google gives | What to check | Fix |
|---|---|---|
| Blocked by robots.txt | A Disallow line matching the sitemap path | Remove the rule, then resubmit |
| HTTP error: 404, or another 4xx code | The address in a private window | Submit the address that exists |
| General HTTP error | The code in the expanded details; Google says a 404 can show here too | Same as a 404, or fix the server for a 5xx |
| Server unavailable | Your host’s status page and error logs | Get the server answering; Google retries for a few days, then stops |
| Manual action | The Manual actions report | Resolve it; Google reads no sitemaps until then |
| Low crawl demand | Nothing on the server is wrong | Google ties crawl demand to content quality: improve the pages, then resubmit |
The second table is “can be read, but has errors”. Google has the file, queues every URL it could parse, and lists the errors on the details page.
| Error | Usual cause | Fix |
|---|---|---|
| Parsing error | An unescaped & ’ ” < or > in a URL | Escape it: & in place of a raw & |
| Unsupported format or Incorrect namespace | A wrong header, a namespace typo, curly quotes from a word processor | Use the namespace http://www.sitemaps.org/schemas/sitemap/0.9 with straight quotes |
| Invalid URL | Spaces, quotes, or a typo such as htp:// | Encode the URL and fix the typo |
| URL not allowed, or Path mismatch | URLs on another host, www against non-www, http against https, or above the sitemap’s folder | Match the sitemap’s host and protocol, and keep the file at the root |
| URLs not followed | Listed URLs that redirect, often several times | List each page’s final address |
| Invalid date | Not in W3C format, or a time with no time zone | 2026-11-16, or 2026-11-16T09:00:00+00:00 |
| Too many URLs, or file size error | Over 50,000 URLs, or over 50MB uncompressed | Split into several sitemaps under a sitemap index |
| Compression error | A broken .gz file | Recompress with gzip and resubmit |
| Empty sitemap | A generator that wrote no URLs | Fix the generator’s output |
| Nested sitemap indexes | An index that lists another index | List sitemap files only |
| Leading whitespace | Blank space before the XML declaration | Google still processes the file; trim it anyway |
| Temporary error | A problem on Google’s side | No resubmit needed; resubmit if it is still there after several hours |
We fed the parse check from the script a one-URL sitemap with a raw ampersand in the address (?a=1&b=2). Python stopped with not well-formed (invalid token): line 2, column 96. With & in its place, the same file printed “parses”. Any URL with a query string carries an ampersand, so check those first.
Which sitemap address to submit on each platform
Google’s help says hosted platforms such as Squarespace or Wix “probably manage your sitemap for you”. Everywhere else, the usual failure is an address the site no longer serves, often after a plugin switch. Each address below comes from the platform’s own documentation, read on November 16, 2026.
| Platform | Address to submit | Note |
|---|---|---|
| WordPress, no SEO plugin | /wp-sitemap.xml | Built in since WordPress 5.5; switched off while “discourage search engines” is ticked |
| Yoast SEO | /sitemap_index.xml | Yoast’s spec has /sitemap.xml redirect here, so submit the index itself |
| Rank Math | /sitemap_index.xml | From Rank Math’s own sitemap guide |
| Shopify | /sitemap.xml | Generated automatically at the domain root |
| Squarespace | /sitemap.xml | On your own domain, or the squarespace.com one |
| Next.js (App Router) | /sitemap.xml | Served from app/sitemap.ts or app/sitemap.xml |
| Astro with @astrojs/sitemap | /sitemap-index.xml | Also writes sitemap-0.xml; our own setup |
The WordPress note catches people at launch. WordPress’s 5.5 announcement says that if Site Visibility is set to discourage search engines, “sitemaps will be disabled”. Untick Settings, then Reading, then “Discourage search engines from indexing this site”, and /wp-sitemap.xml comes back.
Switched SEO plugins? Remove the old entry from the report: click the sitemap, open the more options menu, choose Remove sitemap, then submit the new address. Google’s help says removing a sitemap from the report does not make Google forget the sitemap or the URLs in it.
When the file is fine and the status still says Couldn’t fetch
Google’s help gives two timelines. A submitted sitemap “should be fetched immediately”. After a failed fetch, Google “will retry for a few days, and then stop if the sitemap continues to be unavailable or has critical errors.” The Type column also has a value, Unknown, for a file that “hasn’t been processed yet”. So a new sitemap can sit on Couldn’t fetch while the file loads fine for you.
In that case, work through these in order:
- Confirm the URL Inspection live test shows Page fetch: Successful.
- If a CDN or caching plugin ever served Google an error page at that address, purge it from the cache.
- Resubmit the same address once. Google’s help: “fix the problem and resubmit your sitemap with a new request.”
- Give it a few days before changing anything else.
- Open the Manual actions report, since a manual action is the one cause no request can show.
Low crawl demand is the cause with no switch to flip. Google’s help says “the higher the quality of the site’s content, the higher the crawl demand”, which makes it a content job: fewer thin pages, better ones, and links between them.
After the fix: check the pages it lists
Success means Google read the list, and that is all it means. In Google’s words, “there is no guarantee that a page URL discovered in a sitemap has been or will be crawled or indexed by Google.” The statuses you see next are about pages. Discovered, currently not indexed means Google has the address and has not fetched it. Crawled, currently not indexed means it fetched the page and passed. The Page indexing report can be filtered by sitemap to show which listed pages made it in.
For one page, paste it into our free Google index checker, which reads its robots.txt rule, noindex, redirects, canonical and status code.
For the whole list, an agent saves the clicking. ContextBolt SEO connects your AI agent (Claude, Cursor, Codex, or ChatGPT where your plan allows custom connectors) to Search Console and a site audit. The audit looks for your sitemap on the robots.txt Sitemap: line, then at /sitemap.xml, /sitemap_index.xml and /sitemap-index.xml, follows one level of sitemap index, and checks the most important pages first: 25 by default, at 1 credit a page. A listed page that errors comes back as a broken page or a server error, and one that redirects comes back as “Sitemap URL redirects”, each with its fix.
If the audit finds no sitemap at all, it says so and checks the homepage alone, which tells you the address is wrong before Google does. Search Console reads, URL Inspection included, cost nothing. The full workflow is in a site audit with Claude.
The first task, once you have connected Search Console to your agent:
Run a site audit on example.com. Tell me whether you found a sitemap,
how many pages it lists, and which listed pages error or redirect.
Then inspect the sitemap address in Search Console.
Start the ContextBolt SEO 7-day free trial, which includes 100 credits and needs a card, then paste that prompt. After the trial it is $35 a month for 1,000 credits. How ContextBolt SEO works shows the setup, from one URL in your agent to the first audit.