Short answer: RSS Feed Creator reads a page only when somebody uses it: when you run a preview, create or change a feed, or when a feed is requested after its stored copy has become older than the plan’s refresh interval. Each read is one ordinary request for a public page, with a timeout, a size limit and a user agent that contains the name FeedsBot. It reads public web addresses on the standard web ports and refuses private networks and addresses with a password. If you own a site that people turn into feeds, or you use the tool and wonder what it does to the pages you point it at, this is the whole picture.
Why this is worth knowing
Two groups of people ask this question. The first group are users who want to know whether pointing the tool at a page is polite and predictable. The second are site owners who see an unfamiliar visitor in their logs and want to understand it. Both are served by the same facts, and none of them is a secret: the tool works like a visitor who opens a few pages.
The key point is that RSS Feed Creator does not crawl the web on its own. It never goes looking for pages. Every request it makes can be traced back to a person who pasted an address, or to a reader, a posting tool or a shopping channel that asked for a feed.
When a request is made
| Trigger | What is read |
|---|---|
| You run a preview | The address you pasted, plus what is needed to build the preview |
| You create a feed | The same pages, to build the first version of the feed |
| You save settings or add a source | The feed’s sources, to rebuild the feed at once |
| A reader or tool requests the feed after the stored copy has aged | The feed’s sources, in the background |
The last row is the one that matters for site owners. A feed address is answered from a stored copy. Only if the copy is older than the refresh interval, 3 hours on the free plan and 30 minutes on paid plans, does an update start in the background. So the number of reads of a source is limited by the interval, however many people request the feed. The details are in how often RSS Feed Creator refreshes your feed.
How many pages one update reads
For a feed that is already a feed, an update reads one address. For an ordinary page, it reads the page and, where it helps, a few more. The limits are written into the tool:
- It follows the “next page” links of a list for at most ten pages in total, and stops earlier when the plan’s item limit is reached or a page adds nothing new.
- When more items are needed, it reads the site’s robots file only to find the address of the sitemap, then up to 12 sitemap files.
- For pages found in the sitemap, it reads each page once to get its title, description, picture and date, but never more pages than the feed still needs, which is at most 20 on the free plan and 100 on paid plans.
- Those page details are remembered for about a week, so repeated updates do not read the same pages again.
- Pages are read in small groups, up to eight at a time, not all at once.
In practice, most updates are one request. Section feeds with a sitemap cost more the first time and much less afterwards. This is the behaviour described in RSS Feed Creator beyond page one: whole-section feeds.
A worked estimate
Take a blog with a normal list page, a sitemap and a feed on the PostRSS add-on, which refreshes every 30 minutes. In a day, the feed can be rebuilt at most 48 times, and only if something asks for it often enough. Each rebuild reads the list page, and perhaps a next page or two when the list is short of items. The sitemap files and the individual pages are read on the first build and then served from the one-week memory, so the heavy part happens once. On the free plan, with a refresh every 3 hours, the same feed is rebuilt at most eight times a day.
These are upper limits, not a schedule. A feed that nobody requests is not rebuilt at all, and a feed that is requested all day is rebuilt no more often than its interval allows. For comparison, a person who opens the same list page a few times a day generates a similar pattern of requests, with a good deal more banners and images attached.
Timeouts and size limits
Every request has a leash. A connection that is slow to answer is abandoned after 15 seconds, and a request that takes longer than 25 seconds in total is stopped. Large responses are cut at a size limit, which is several megabytes for a normal page, so a page that is huge or never ends cannot tie up the tool. Redirects are followed, a few steps at most.
For you as a user, the practical result is that very slow sites can fail. When that happens, the update does not break your feed: the previous version stays in place, and the management page shows a “Last update” line that explains what happened. The behaviour is described in when a page layout changes: how RSS Feed Creator copes.
How the tool identifies itself
RSS Feed Creator does not pretend to be a person with a browser. Its user agent begins like a normal browser string, for compatibility, and then says it is FeedsBot. A site owner can search the access log for that word to see which pages were read and when. It also sends a request header that prefers English content, which helps with sites that serve several languages.
This matters because a site that wants to understand its traffic can: the visits are tied to the name, they come in small groups, and they follow the pattern described above. There is no hidden reading under another name.
Which addresses are refused
Before anything is read, the address is checked. The tool accepts only these:
- web addresses that start with http or https, typed with or without the prefix;
- the standard ports 80 and 443;
- real domain names, not bare numbers;
- addresses with no username or password inside them;
- names that point to public internet addresses.
Addresses that point to private or internal networks are refused, and the same check applies after redirects, so a public address cannot lead the tool into a private one. This protects everybody, including you if you run the tool against your own sites, and it is also why pages behind a login cannot be turned into feeds. A related protection applies to the preview: an address the tool cannot read produces a clear message, not a half-built feed. The empty-preview case has its own guide: preview shows 0 items: what to try.
What the feed itself looks like to the outside
The finished feed is served from RSS Feed Creator, not from the source site. The address ends in .xml, the response may be cached by readers for a few minutes, and the feed carries an instruction that asks search engines not to index it. That keeps a copied list of headlines from competing with the original site in search results.
The items in the feed point to the original pages. Links go to the source site, and a picture in the feed is only an address: the file stays on the original website, as explained in pictures in your feed: where RSS Feed Creator finds images. Readers who click go to the source, which is the way a feed is meant to work.
A short checklist for users
- Choose the page that lists items, so that one request gives you most of what you need.
- Prefer a feed address when the site has one; it is the lightest read.
- Use filters in the tool instead of many near-identical feeds of the same page.
- Do not create feeds you no longer use; the management page has a delete button, and the reads counter shows whether anything still fetches the feed.
- Check the preview before you create the feed, because it costs nothing and shows what will be read.
The plans differ in limits and refresh intervals, not in how pages are read; see the pricing page for the numbers.
Related reading
- Is It Legal to Create RSS Feeds from Other Websites?
- How RSS Feed Creator Finds Items on Pages Without RSS
- A Tour of the RSS Feed Creator Management Page
The bottom line
RSS Feed Creator reads on demand, in small groups, with time limits and a name. Most updates are a single request, the refresh interval caps how often a source is read, and private networks and logins are off limits. That makes the tool predictable for the people who use it and recognisable for the sites that are read.
BUJ
Does the tool read pages in the background all day?
No. A page is read when you preview, create or change a feed, or when a requested feed is older than its refresh interval.
How can a site owner recognise the reads?
The user agent contains the name FeedsBot, so the access log can be searched for it.
Can the tool read a page that needs a password?
No. It reads public pages only and refuses addresses with a username and password.
What happens if a page is slow or fails?
The update is abandoned after a time limit, the previous version of the feed stays in place and the management page shows what happened.


