ClipAI - Knowledge Base (FAQ)
< All topics

How to speed up the program?

Are articles being generated too slowly? There are several ways to speed things up:

1. Multi-threaded mode

The most effective way to speed up. To use it, in the “Input Data” tab, select multi-threaded mode, and then at the bottom of the Zennoposter window, in the “Settings” tab, enter the number of threads in “Maximum threads”, and in “How many to do” - the number of lines in the file that need to be used for generation (you can enter more or click on at the top of the Zennoposter window).

1 thread = 1 concurrently running copy of the program.

2. Changing the rewriting method

“Rewriting” (text synonymization) is the most labor-intensive and time-consuming way of working due to the nature of AI. In some situations, it is more appropriate not to rewrite the text, but to write a new one in the “Writing an article” mode. If this is not possible, try switching the rewriting uniqueness to “Medium” in the “Processing” tab - this should speed up the rewriting several times.

3. Reduce the amount of text generated

In the "Processing" tab, you can specify how many words or characters you need in the text. Perhaps you need smaller text? Then there will be much fewer requests, which means the text will be generated much faster.

4. Disable unnecessary options

Almost everything affects the speed of text creation, for example:

  1. Language: Cyrillic text takes approximately 2 times longer to create than Latin text;
  2. HTML markup usually speeds up generation a little on small text, but slows down on large text;
  3. Creating images for text significantly slows down its publication;
  4. In the “Parameters” tab of the input settings, with the “Priority on...” - “High text quality” parameter, articles are generated longer than when using “Token Saving”.
  5. The mode of writing short articles (without subheadings) is much faster than the mode of writing articles according to plan. However, the short article mode can work much longer if you generate more than 4000 characters in it, while using the “Check for matches when continuing text” option in the “Processing” tab.