Supertweet

I have a bad habit of firing off a post and deleting it twenty minutes later. Rather than try to fix that with willpower, I built the gate.

Supertweet scores a draft against criteria you define once — tone, whether there is an actual claim in there, whether it reads as a jab at a named person — and then replays it through several audience perspectives, which turns out to be more useful than the score. The same post read as a hiring manager and as a peer in the field are frequently two different posts, and you only notice after you have sent it.

It also scores against what tends to earn reach, which is a heuristic built from observed patterns rather than any privileged knowledge of a ranking system.

It also works as a linter: point it at a draft and it flags the specific lines that trip your own rules, rather than handing back a verdict on the whole thing. A score tells you something is off; a line number tells you what to change.

Chrome extension and web app, both shipped in a day. It works on X and LinkedIn.

It will also refuse to send. Below a threshold you set, the post is locked and the panel shows what would lift it — each suggestion with the points it is worth, so the fix is a specific action rather than "make it better."

The screenshot is the tool grading a post about itself. It scored 5.0 and locked it, on the grounds that the draft read like a feature list with no screenshot, no result and nothing learned from building it. That was correct, and it is the reason there is a screenshot on this page.

The part doing the real work is not the model. It is the pause.