Troubleshooting and log

Most problems can be solved with the Syndicator log, which records every step of the last feed run, and a manual run with Pull selected feeds now!, which shows the same report on screen.

The Syndicator log

Open RSS Retriever › Syndicator log. The log covers the most recent run: every item, every step and every error, with a time stamp. A new run replaces the previous log. Click Copy to clipboard to share it on the support forum, after removing anything private.

A typical successful item looks like this:

[06-10-26 09:15:02] Import https://news.example.com/bike-lanes
[06-10-26 09:15:03] Extract full text article
[06-10-26 09:15:04] Done
[06-10-26 09:15:04] Process post templates
[06-10-26 09:15:04] Generating content with Claude (claude-sonnet-4-6) for: Rewrite ...
[06-10-26 09:15:31] Success

The feed does not load

Message: Unable to acquire https://..., or The source feed is empty.

Open the URL in your browser. If it shows an HTML page instead of XML, look for the real feed link on that page (often in the page source as application/rss+xml).
The site may block scripts. Set a browser User agent, an HTTP referrer or HTTP headers; for a new feed use Quick override default settings.
Your server IP may be blocked. Use the proxy list and keep Use Wayback Machine on.
XML parse error: the feed is not valid XML. Try UTF-8 encoding or Convert character encoding.

Feeds are not updated automatically

Check Next update in the feed list and the Check for updates every value; 0 disables automatic updates.
In Auto pull mode, WordPress cron runs only when the site has visitors, and some hosts disable WP-Cron. Switch to Cron job or manually and add the cron command shown on the General Settings page.
With a cron job, make sure your caching plugin or CDN does not cache the ?pull-feeds= URL.
Runs that stop halfway usually hit the PHP time limit. Raise Max execution time or import fewer posts per update.

Items are skipped

Log message	Reason and fix
`The post already exists ID: ...`	A duplicate. See Check for duplicates by. Posts in the trash do not count; deleted posts that are still in the trash do.
`No mandatory keywords or phrases found in the post`	The item did not pass a keyword rule. Keywords match whole words; add other word forms.
`The post contains at least one of blacklisted keywords or phrases`	A word from Must contain none of these keywords was found.
`The post is too short` / `The post is too long`	The length rules. Feeds with excerpts only are often too short: turn on full-text extraction.
`No mandatory categories found in the post.`	Undefined categories is set to import only items that match your categories.
`The post is empty`	Neither the feed nor the templates produced content. Check the Post content template.

Remember the date rules and Maximum number of posts: a feed may simply have no new items that qualify.

Full-text extraction fails

Message: Operation failed. Unable to retrieve full-text content from ... followed by The post will not be added.

Look at the Full-text article preview. If it is empty, the method does not work for this site.
With the Full-Text RSS script: switch to custom settings and name the container element of the article.
With custom settings: check the tag and the JSON attributes against the HTML of a current article. Sites change their layouts.
The site may block the request; see The feed does not load.

AI requests fail

Log message	Reason and fix
`OpenAI API key is not set` (and the same for other services)	Enter the key on the Accounts page. Check that the engine name points to the provider you have a key for; unknown names go to OpenRouter.
`... hourly request limit has been reached`	The request limit you set for that service is used up for the current period. Raise it or wait for the next period.
`Error: ...` with a message from the provider	The provider refused the request: invalid key, no credit, unknown model name, content policy, or a rate limit. The text after “Error:” comes from the provider and names the cause.
`Rate limit reached ... Please try again in 20s`	Too many requests per minute for your provider account. Set a Delay and lower the number of posts per update.
`[gpt_article] was unable to generate a N-section list`	The model did not return a usable outline. Try another model, simplify `structure`, or reduce `sections`.

When an AI shortcode fails, the whole post is skipped rather than published half-finished. The next run tries the item again, as it has not been imported.

The answer contains Markdown such as **bold** or ###? Tick Convert Markdown to HTML, and ask for HTML in the prompt.

Translation fails

Message: Translation failed. The post is not added. Check the API key and the remaining character quota of your translation service, and for DeepL whether Use DeepL API Free matches your key type (Free or Pro).

Images or featured images are missing

The thumbnail source image was not found or broken: the source of Post thumbnails has no usable image. Add an Alternative thumbnail source.
The thumbnail source image does not match the specified size limits: relax the Thumbnail source size values or set them to 0.
Unable to save image to disk: the uploads folder is not writable, or the GD extension is missing.
Images are broken on your site: the source blocks hotlinking. Tick Store images locally and turn off FIFU.
Unable to retrieve Pixabay image / The requested image was not found on Google Images: use simpler, English keywords.
An image generator reports an empty prompt: check that the prompt field is filled and that its placeholders return text.

Server requirements

A red notice on the RSS Retriever page names a missing PHP extension (mbstring, GD) or a disabled allow_url_fopen. Ask your host to enable it. See Requirements.

Getting help

Ask on the RSS Retriever support forum. Include the feed URL, the relevant part of the Syndicator log and the settings you changed. Never post API keys or your registration data.