Test your tweet before you post it
Paste a draft and get a viral score in a second. Then let eight AI writers rewrite it, round after round, until it beats your normal post.
How it works
Most viral scores grade you against everyone. This one starts with you: your posts, your audience, your normal.
- 01
Score
An AI evaluation model (Jev) answers 56 yes/no questions about your draft in about half a second: does it make the reader feel something, is there a claim to argue with, is the first line strong? Eight signals, one score.
- 02
Calibrate
Add your handle and we read your last 60 posts the same way, then learn which signals actually move your likes, replies and reposts. From then on 50 means your normal post, not anyone else’s.
- 03
Rewrite
Eight AI writers (GPT-5.5, Claude Sonnet 5, Grok 4.6 and GPT-5.6) each rewrite the post their own way. Every version is re-scored and checked for faithfulness: same point, nothing invented. The best go back for another round.
Watch a real run →
Questions, answered plainly
Isn’t a viral score just a guess?
It’s a comparison, not a forecast. The model answers 56 yes/no questions about the text; each of 8 signals is compared with a baseline (your own normal post, when you add a handle) and weighted by what that audience rewards. It reads the text, not the algorithm: timing, images and who sees the post first still matter.
Will the rewrite change what I’m saying?
It must not, and it’s checked. Every rewrite is scored for faithfulness against your original: does it keep your point, does it invent facts, experiences, advice or villains? Rewrites that drift are shown crossed out and can never win. A borderline one is flagged as a close call.
Why only one rewrite run a day?
Each run sends eight AI writers through up to five rounds and scores every version, which costs real money. So it’s one free run per visitor per day. Live scoring keeps working without limits.
What do you store?
Scores are cached for 30 days under a hash of the text, never the text itself. Rewrite runs keep their text for 90 days so the replay link works. Nothing is posted for you, and there is no sign-in: you type a handle, never a password.
Can I test any handle, or none at all?
Any public account. With no handle, or fewer than five original posts, drafts are scored against a typical post on X and the score says so.
Who built this, and what’s the catch?
Octolens, an AI social listening tool. It reads every mention of your keywords across 15+ platforms and scores each one for relevance: the same read-and-score loop, pointed at the whole internet. No catch: the tester is free and needs no account.
Built by Octolens
Post it where the conversation already is.
This tool reads one post and scores it. Octolens does the same for every mention of your keywords across X, Reddit, LinkedIn, Hacker News and 10+ more platforms, with AI relevance scoring, sentiment and Slack alerts. Better content starts with knowing which conversations to join.