Substack keyword parsing not working

[Pseud3mys]

Hello, I am new in writing Github issu, sorry for the probable unusual style.

I added Substack RSS feeds to the app and configured the gemini api key. But the only key words generated are: "substack.com", "jptorrents.substack.com". And in one of the post I have "upper-case-title". Maybe the bot can't access the article but only the title... I think the issu can be encontered on other news web sites and it make the app unusable :'(

Great project by the way ! Thank you.

[Tiendil]

Hi!

Thank you for the report and for the good words!

A few questions to clarify the issue:

• What action did you perform first: added the feed or set the API key?
• Were there new posts in the feed after you set the API key?

Here is some info about how feeds.fun works currently:

• There are multiple tag processors that process each post from each feed in the order they are received.
• Here are some of them: a processor that sets tags according to the domain, a processor that detects UPPER CASE TITLES, and LLM processors.
• If a processor can not process a post, for example, there are no users with API keys subscribed to the feed, it will skip the post and it will never return to it.

If you see the tags like substack.com, jptorrents.substack.com, upper-case-title, then I assume that the feed was loaded correctly and the tag processors processed its news.

=> If you already set the correct API key, the LLM processor should process the next news after that moment, and you will see tags for it.

P.S. Also, LLM processors have a limit on the age (time of publishing) of processed posts. It was made to avoid overspending on processing too many old posts. For example, if a feed will return 10000 posts. You can find this limit in the settings. So, even if you add the new feed after you set the API key, not all posts from it will be processed — only the recent ones.