Is it possible to send single request for rewriter? I am using Gemini API and 1minAI’s API, Gemini has limit of 20 RPD (I am using almost 10 free APIs), as for 1min.AI’s API, it has limit of 180 RPM, Also, when rewriter send multiple requests to rewriter content, it often gets error after some requests and token consumption is become too high due to multiple requests.
Is it possible to send the whole article in single request? I am using Google AI Studio’s app (build myself) to humanize the content and it is keeping the score of content under 30% in zerogpt.com, I have tried same with 1min.AI chat module too. Both were giving good results with same prompt. But it is too much time consuming and hassle to do this for each article separately. SCM is the best in this regard, only problem is that it send requests by paragraphs which often causes issue with APIs.
If it can send single requests with whole article and instructions, it will solve the API requests issue.
I tried doing same with Page Pipeline tool, but it was not picking existing articles at all. So, it was not useful in this case at all.
Yes, I am trying to rewrite an AI Article, first I generate the article using Page Pipeline tool, Then I use prompt with Gemini-2.5-flash/GPT4.1 to humanize this article. I wanted to bulk rewrite articles using the prompt.
As for article loading in page pipeline too, I have added its variables %scraped_article% and <load_next_article> from variables from SCM, but none were picking articles from folder. To use those article files as reference, I had to downgrade the SCM to 11-2025 version, that was picking articles automatically. But after that no version picked the articles from folder.
I am giving SCM folder of articles to rewrite using Rewriter tool but it is only taking paragraphs, and sending requests, which fills the API quota and sending multiple requests in short time is causing requests failures too. It oftens trigger RPM and failure of requests too.
Ill have to check if those macros are broken then.
Yes folder insert is not the way, it will chop up your content and spin per sentence.
Issue was local article parsing.
Some source files were valid HTML fragments but had no <body> tag.
Parser expected a body node, so it could fail after finding the <h1> title.
That made %article_title% and %article_body% blank even though file was found.
Fix added:
Parser now supports HTML fragments without <body>.
Local article import now falls back to raw <h1> / filename and text body if parser fails.
Logs now show loaded article title/body size, plus fallback notice if used.