Research
Gathers the material the article will be built from, so the draft has something to say beyond what a title implies.
WPCWP20 — until 31 October 2026
Home / WordPress AI content generator
WP Content Writer Pro generates drafts in your own WordPress install, using an API key you supply. The output is an ordinary post, and nothing publishes itself unless you ask it to.
A run ends with a row in wp_posts: a normal WordPress post with a title, body, excerpt, categories and a status. Your editor opens it, your SEO plugin reads it, your revision history covers it.
That matters more than it sounds. A generator that stores content in its own tables, or renders it through a shortcode at page load, takes your archive with it when you deactivate it. This one does not, and it is worth testing that claim on a staging site rather than believing it.
Gathers the material the article will be built from, so the draft has something to say beyond what a title implies.
Decides the shape before any prose exists. Most of the quality of the finished piece is settled here.
Writes the body against that outline. This is where most of the token cost sits.
Title, excerpt and meta description, then a pass over the prose. Each stage is a separate call, so cheaper models can do the cheaper jobs.
The default post status is draft. Nothing reaches your front page without a person opening it first.
The fact-checking pass marks what it could not verify instead of quietly rewriting it, so you can check the specific sentence.
Optional, through Copyscape or Winston, using your own account with either.
Input and output tokens per call with an estimated cost, attributable to the article that caused them.
Generation is billed by your AI provider, at their rates, against your key. We charge for the software per site and take no cut of the tokens.
Two things move the number more than people expect: which model runs which stage, and whether failed attempts are counted. A call that consumed tokens before erroring still costs money, so a usage log that omits those will not match your provider's invoice.
The cost calculator works out a per-article figure for each supported model.
Only if you change the setting. The default status is draft, deliberately: an unreviewed article going live under your name is a much worse failure than a review queue is an inconvenience. Direct publishing exists, but you have to go and turn it on.
Nothing. Each run writes an ordinary WordPress post, so drafts and published articles stay exactly where they are, editable in the normal editor. There are no shortcodes to leave behind and nothing is rendered through the plugin at page load.
You pay your AI provider for the tokens and nothing to us for generation, so the figure depends on the model you pick and the length you ask for. The cost calculator works it out per model using the same rate table the plugin bills its own usage log against.
Google's stated position is that it rewards helpful content however it was produced, and acts against low-value content produced at scale to manipulate rankings. That makes it a question about your editing process rather than your tooling.
The calculator uses the same rate table the plugin bills its own usage log against, so the two agree by construction.