v1.2.0
New
- The AI-likelihood check now runs in your browser: each visible post gets an AI N% from common AI-writing patterns, with no key and nothing sent, and clicking it lists the patterns behind it.
- You can hide posts at or above an AI-likelihood you choose. Each hidden post has a Show post button, and hidden posts stay out of Top posts and Find leads.
Improved
- The local check now recognizes casual AI-written replies: a stock reply frame, a polished lowercase one-liner, a laugh tacked onto a clean sentence, or a generic yes-or-no question to the author. Reply signals alone stay at or below 65%, under the default hide threshold.
- One pattern alone is capped well below the hide threshold, and a very short post with no clear sign shows no verdict.
- With the TypeSafe score also on, Jev's rating is blended with the pattern score, and the question sent to Jev no longer leans toward a human writer for short posts: length alone is not evidence either way. The same fields go to the same place, with your key.
Fixed
- Security advisories in the development tools are patched. The libraries bundled into the extension are unchanged.
Action needed
- If you had turned on AI-likelihood settings while the TypeSafe score was off, a one-time update turned them off. Turn the check back on in the popup if you want it.
v1.1.0
New
- Pro plans through ExtensionPay on Stripe: $4.99 per month, $29.99 per year or $49.99 once. Scoring and the TypeSafe score stay free.
- Pro product promotion: scan your product's landing page, find posts on x.com where it helps, and get reply drafts from your own AI key that you edit and post yourself.
- The Top posts panel (Pro) ranks the current page's top 10 posts by Jev (TypeSafe), Algorithm or Combined, with each post's scores, author, followers and posted time.
- Author follower counts (shown on badges with Pro), so a new post from a large account can rank before it gathers likes.
- An optional TypeSafe score beside the local score, using your own key. Off by default.
- Keyword learning (Pro) from the newest 50 posts you liked, replied to or bookmarked, plus your profile. You review every change and can revert the last one.
- Mute keywords, and a learning reset.
- A Settings tab for API keys, with Test buttons, and for the reply style.
Improved
- Scoring now targets how likely your reply is to be seen: it weighs how fast a post is gaining likes and retweets for its age, keeps posts with almost no engagement out of High, and pushes down crowded reply threads.
- API keys are kept in the extension's own storage, which x.com pages cannot read. Existing TypeSafe keys move there on update.
- The Top posts panel's System, Light or Dark theme is now set in the popup's Settings tab and applies at once.
Fixed
- Higher contrast in the popup and on score badges.
- Badges repaint or clear after every settings save.
- Closing or leaving a tab while badges update no longer causes errors.
v1.0.0
New
- First release of X Post Radar: scores and highlights the posts on your X timeline.
Improved
- Faster on long timelines: only newly added posts are checked, not the whole page.
- Clearer popup controls and larger buttons for removing keywords.
- The popup no longer loads Google Fonts.
Fixed
- The popup checks threshold values and shows storage errors.
- Settings are saved before posts are rescored.
- Skipping a post after engaging with it no longer counts as a skip for learning.