A simple tool to convert an XML sitemap into an Atom feed — especially useful for sites that don't have a CMS, or where the CMS doesn't produce a feed. Each URL in the sitemap is fetched and its OpenGraph and Twitter Card metadata (title, description, image, author, dates) is used to build a rich Atom entry.
Once published to PyPI you can run it directly with uv:
uvx sitemap2atom https://example.com/sitemap.xml -o feed.atomTo run the latest code straight from GitHub (before a release, or to try main):
uvx --from git+https://github.com/darkflib/sitemap2atom sitemap2atom https://example.com/sitemap.xmluv tool install sitemap2atom # installs the `sitemap2atom` command
# or
pip install sitemap2atomsitemap2atom SITEMAP_URL [OPTIONS]By default the feed is written to standard output; redirect it or use -o to
save it to a file:
# Print to stdout
sitemap2atom https://example.com/sitemap.xml
# Write to a file, limiting to the first 20 URLs
sitemap2atom https://example.com/sitemap.xml -o feed.atom --limit 20-o, --output PATH— write the Atom feed to this file (default: stdout).--limit N— maximum number of sitemap URLs to process (default: all).--feed-title TEXT— title for the generated feed (default:Enriched URL Feed).--timeout SECONDS— per-request timeout in seconds (default:10).-v, --verbose— enable info-level logging on stderr.--version— show the version and exit.
from sitemap2atom import fetch_sitemap_urls, enrich_url_list_to_atom, feed_to_pretty_xml
urls = fetch_sitemap_urls("https://example.com/sitemap.xml")
feed = enrich_url_list_to_atom(urls[:10], feed_title="My Feed")
print(feed_to_pretty_xml(feed))See this gist for a sample of the kind of enriched Atom feed produced: https://gist.github.com/Darkflib/989b8f3a5a1ea995e8e294669d5e282a
This is a simple tool aimed at basic use cases. It does not support authentication, sitemap index files / pagination, or dynamic sitemaps, and may not handle every sitemap or page format. Treat the sitemap and the pages it references as untrusted input and run it against sources you trust.
Some sites sit behind bot-protection that serves a JavaScript "verify your device" challenge instead of the real content. sitemap2atom sends browser-like headers, which is enough for many of these, but sites that require JavaScript execution cannot be fetched by a simple HTTP client. In that case you'll see a clear error explaining that an HTML page was returned instead of a sitemap.
This project uses uv.
git clone https://github.com/darkflib/sitemap2atom.git
cd sitemap2atom
uv sync
uv run pytestSee CONTRIBUTING.md for more, and CHANGELOG.md for release notes.
Editorial pipelines: Sites without native feeds (static sites, marketing pages) can now be consumed as structured feeds, ready for newsletter aggregation or content curation workflows.
Agent pipelines: LLM agents and agentic systems need structured, on-demand inputs. sitemap2atom turns any public web site into machine-readable content without requiring API keys or custom integrations.
This tool is maintained by Mike Preston, who consults on agentic infrastructure at wwff.tech.
This project is licensed under the MIT License — see the LICENSE file for details.
PS. If you do anything interesting with this code, please let me know! I'd love to hear about it.