The Publish step of every post opens with a quality check card: the post's score out of 100, how each of the five areas is doing, how many sources it cites, how long it is, and how hard it is to read compared with the articles that already rank for its keyword. It is the same card PixelClaw puts on its own articles, filled with your post's real numbers. It adds no new score: the number is the one you see in Optimize and in the checks panel.
Find the card
Open your Blog Site and choose a post.
Open the Publish step.
The card sits at the top of SEO & AI readiness. Under it you find a one-line verdict, Open the Optimize step, Open the checks, and the publish controls.
The card describes the last saved version of the post. After you edit and save in Draft, reopen the post to see the new numbers.
What each part of the card means
The small line at the top reads PixelClaw quality check • Passed when the score is 60 or more, • Not yet when it is lower, and • Not run before the post has been saved and scored.
The title says 5-area quality bar passed (72/100) or Quality bar not passed yet (54/100). The number is the post's quality score, rounded.
The badge on the right says Publish bar passed at 60 or more, Autopilot passed when a post Autopilot wrote (including God Mode's queue) scores 77 or more, and Below the publish bar under 60.
Citations counts the sources listed in the post's Sources blocks. Each address counts once, however often it appears.
The five tiles show the share of each area's applicable checks that pass, with the count underneath, such as 7 of 9 checks. They are the same five areas as the checks panel and the score: SEO (Search) is Search engines, AEO (Answers) is Answer engines, AIO (AI Overviews) is Trust signals, SXO (Experience) is Reader experience, and GEO (Discovery) is AI discovery.
Readability shows the post's reading grade, and the grade of the top ranking articles when PixelClaw has measured them for this post, for example Grade 13 · top articles read at grade 12. The sections below explain both numbers.
Word Count is the post's body length.
The last line says Built & scored using PixelClaw Autopilot for a post Autopilot wrote, and Scored by PixelClaw for a post a person started.
What "passed" means for publishing
Publishing by hand needs a score of 60. An owner or admin can still choose Publish anyway below the card; the choice is recorded in the Activity log with their name. Autopilot publishes on its own only at 77 or more, and otherwise waits for your review. So an Autopilot post between 60 and 76 shows Publish bar passed: you can publish it yourself, but Autopilot would not have.
Passing the bar does not mean every check passed. Open the checks for the findings that remain, and read the article before you publish it. Understand your content score and the 27 checks explains each check and how the five areas make one score.
How the reading level is measured
PixelClaw uses the Flesch-Kincaid grade, the standard measure behind most readability tools, on the exact text of the post. It counts the words, the sentences and the syllables, then combines the average sentence length and the average syllables per word into a US school grade. Grade 8 means a typical 13- to 14-year-old can read it comfortably; grade 12 is the end of high school; 13 to 16 is college level; 17 and above is graduate level.
The count covers what a visitor reads: headings, paragraphs, lists, callouts, key takeaways, FAQ answers and table cells. Images, code and the Sources list are left out. The measure applies to English posts of at least 200 words; shorter posts and translations show no grade. The card shows the exact grade, rounded to a whole number.
Why the card compares it with the top articles
There is no single right reading level. A guide to container networking read by engineers sits at college level in every article that ranks for it; a recipe does not. Judging both against one fixed bar would be wrong for one of them.
So PixelClaw measures the three articles that rank for the post's keyword, the same three that set its planned length, and takes the middle grade. The writer is told to write at that level, and the card shows it next to your post's own grade. Reading within a grade or two of the top articles is the goal. Well above them means your post is harder to read than readers of that search expect; well below them can read as thin to an expert audience. A technical post at college level is right when the ranking articles read that way too.
Readability is a fact on the card, not a check. It never changes the score and never holds a post back.
When PixelClaw has no measured top articles for a post, the card shows your grade alone. That is the case for posts researched before November 8, 2026, posts written without research, and imported posts.
The reading level on your blog
Visitors see the level in the post header, next to the reading time: 5 min read · Grade 10 reading level. It uses bands rather than a verdict: Grade N reading level up to grade 12, College reading level for grades 13 to 16, and Graduate reading level from 17. It appears on English posts of at least 200 words; translations and shorter posts show the reading time only. It is worked out from the published text each time you publish, so an edit that simplifies a post updates it.
Bring a post closer to the top articles' level
Readability has no one-click fix, because it is not a check. Edit in Draft, save, and publish:
Split long sentences. One idea per sentence brings the grade down fastest.
Use plain words where the subject allows, and keep the terms your readers use. Replacing a term your audience searches for makes the post worse, not better.
Turn a dense paragraph into a list or a table when it holds steps or comparisons.
To raise a grade for an expert audience, add precision rather than long words: the exact term, the number, the condition under which the advice holds.
Related guides
Checked against PixelClaw's scoring and readability rules on November 8, 2026.