How do you find blog topics from RSS feeds and Google Trends?
In short: RSS feeds give you the new posts on sites you follow, as a title and a short summary; Google Trends' daily feed lists the searches that are rising in a country that day. Read both, keep what fits your site, and you have fresh topic ideas every day. The hard part is not writing the same topic twice under a different title. Comparing headlines and URLs isn't enough for that; you also have to compare each new topic with the posts already on your site.
Why are RSS feeds useful for finding topics?
Most news sites and blogs publish an RSS feed. On WordPress sites you can usually open it by adding /feed/ to the address. Each entry has a title, a short summary and a link. That's enough to see what's being talked about, without reading or copying the full article.
Wiozen Content Bot reads only the title and a summary of up to 45 words from each feed entry. It never fetches the article itself. The post is written from scratch starting from that headline, and the source is credited at the end.
You don't have to hunt for feed addresses one by one. Press Find sources automatically and the model suggests up to 12 sites that fit the category brief, with two web searches adding more. The home page and /feed/ address of each site are actually fetched, and only feeds that return at least one entry are added. It can take a minute or two and uses a few API requests.
What does Google Trends give you, and what doesn't it?
Google Trends' daily trending feed lists the searches that are rising in the country you pick. Each term often comes with a related news headline, which the bot adds as context. It takes up to 20 terms from the feed per run.
Trending terms are broad: a match, a TV series, an election result. Most of them won't fit a site about recipes or loans. So Trends makes sense for news-style categories. In the Content Bot each category has its own Trends checkbox, and the country is set as a two-letter code on the Connection and automation tab, TR by default.
How do candidates turn into topics?
For each category the bot collects up to 120 candidates per run. That limit is shared across the feeds, 2 to 12 entries each, with 20 slots kept for Trends when it's on. Reading feeds gets a 60-second budget and the order is shuffled every run, so one slow feed doesn't keep the same sources out every time. WordPress normally caches feeds for 12 hours; the bot cuts that to 30 minutes so the news stays current.
The candidates then go to the model together with the category brief. The model keeps the ones that really fit and writes, for each, a new headline that doesn't copy the source's, a focus keyword and the angle the article will take. If nothing fits, it returns an empty list instead of forcing a match. The bot then suggests a topic itself based on the brief, so a category doesn't go empty on a quiet news day.
How do you avoid writing the same topic twice?
Repeats are caught at two levels: the same news item showing up again, and the same subject being suggested again under another headline.
The same news item
Each candidate gets two fingerprints. The title fingerprint ignores letter case, the Turkish dotted and dotless i, punctuation and spacing. The URL fingerprint ignores the scheme, www, a trailing slash, the page anchor and tracking parameters such as utm_, fbclid and gclid. If either has been seen before, the candidate is dropped, so a story whose headline was edited is still caught as long as its URL stays the same. The record is kept in two places: the last 3,000 marked topics, and hidden fields on the posts the bot wrote. The second lasts as long as the post exists.
The same subject, a new headline
Because the model writes a fresh headline every time, fingerprints aren't enough on their own. With "Gavurdağı Salad Recipe: Walnut Meze" already on the site, the model might suggest "Gavurdağı Salad with Walnuts". So before anything is written, each chosen topic is compared with every post on the site (published, draft and scheduled):
- The keys are the part of the title before the first colon, pipe, bracket or dash, and the focus keyword stored by the SEO plugin.
- Words that appear in at least a fifth of the site's posts ("recipe", "how to") count as boilerplate and are left out. This list isn't built per language; it comes from the site's own titles.
- A topic whose title matches an existing one word for word, boilerplate aside, is dropped straight away. A topic that shares word stems with existing posts is sent to the model along with up to five of the closest posts and the question "does this answer the same question?". That way early repayment on a mortgage and early repayment on a car loan count as different topics, while a real repeat is dropped.
- Every skipped topic is written to the log along with the post it matched.
News topics (from RSS and Trends) are only compared with posts from the last 30 days, so a story that comes back every month, like an interest rate decision, can still be covered. Topics the bot suggests itself are compared with no time limit. When suggesting, the model sees up to 600 existing topics, those in the same category first. If every suggestion is already on the site, it's shown what was rejected and asked again, up to four more times; in those rounds it splits the category into narrower subtopics and looks for gaps.
What the check can miss
If two sites cover the same event under different headlines and URLs, and the two headlines share no word stems, both can come through as separate candidates. Because of the 30-day window, a topic from RSS can be written again a month later under a different title. And if the model doesn't answer the comparison question, a stricter word rule takes over, which now and then drops a topic that was actually different.
The Log tab is the quickest way to spot these. Each "Topic skipped" line says which existing post caused the skip. If you see good topics being dropped, a narrower category brief usually helps.
Frequently asked questions
Can I copy a news story from an RSS feed onto my site?
Copying the full article raises copyright problems. Use the feed to find topics and write the text yourself. The Content Bot reads only the title and a short summary from the feed, writes the post from scratch and credits the source at the end.
How many RSS feeds can one category have?
Up to 40. In each run up to 120 candidates in total are read from those feeds and Trends.
Which country does Google Trends use?
You enter a two-letter country code in the settings (TR, US, DE and so on), and the bot reads that country's daily trending feed. The default is TR.
If a headline changes, will the bot write the story again?
Not while the URL stays the same; the URL fingerprint ignores tracking parameters and small formatting differences. If both the headline and the URL change, the topic-level comparison takes over.
What happens if the sources have nothing suitable?
The bot suggests a topic from the category brief and checks that suggestion against the site's posts too, so every active category still gets an article each run.
To try it on your own site, download the Content Bot, add a few feeds to one category and check the log after a first run with Run now. Setup steps are in the WordPress auto-publishing guide.
More guides: How can you use an AI API from Turkey without a credit card? · How do you connect an OpenAI-compatible API to a tool you already use? · How do you calculate the cost of an AI API? · How do you get AI to write and publish articles in WordPress automatically? · Does AI-written content show up on Google and in AI answers?