Summary

This free ai rewriter scores a cold email or DM out of 100 against eight writing rules, then applies fixed text rules to produce a cleaner version. It is not an AI model: it removes pleasantry openers, swaps stiff phrases, drops hedging words and splits long sentences, and it flags what needs a human edit. Everything runs in your browser, and nothing is uploaded.

AI Rewriter for Cold Messages, Without the AI

Paste a cold email or DM. Get a score out of 100, eight pass-or-fix checks and a cleaner version. It runs fixed rules in your browser, not a model.

Try it on your message

Cold message checker and rewriter

Paste your draft or load the example. The score and the rewrite update as you type.

0 words
0/100
Paste a message to score it.

Checks (8 rules, 12.5 points each)

    Rule-based rewrite

    No AI model runs here. The rewrite applies fixed text rules in your browser, and nothing leaves this page.

      How it works

      What the score counts, and what the rewrite changes

      Eight rules, 12.5 points each

      Length, opener, reader focus, buzzwords, sentence length, question count, hedging and tone. Each rule is pass or fail. The score is passed rules divided by eight, times 100. No weighting, no hidden model.

      A rewrite built from text rules

      It removes pleasantry openers, replaces phrases like "reach out" and "leverage" with plain verbs, drops "just" and "perhaps", turns exclamation marks into periods and splits sentences over 25 words at ", and".

      A list of what it left to you

      Self-introductions, buzzwords, extra questions and messages over 120 words are flagged, not rewritten. A rule cannot invent a fact about your reader, so we do not pretend it can.

      Use it in 3 steps

      From draft to sendable in about 2 minutes

      1. 1

        Paste the draft as you wrote it

        Include the greeting and sign-off. The tool skips them when it counts the first sentence.

      2. 2

        Fix the red items first

        Start with the opener and the question count. Those two change how the message reads in the preview line of an inbox.

      3. 3

        Replace the generic line with a real signal

        Quote the post, the date and the platform. Name, post, date: a named person with a dated post beats any polished opener.

      Why rules, not a model

      A rewrite you can predict and audit

      A language model rewrites differently every time and can add claims you never made. A rule list does not. Run the same message twice and you get the same output, and each change is listed under the rewrite so you can reject it. For outreach, where one invented claim can sink a thread, that predictability matters more than flair.

      • Same input, same output, every run
      • Every change is listed in plain words
      • Nothing is uploaded or stored
      • Nothing is added that you did not write
      Hand circling a single word on a printed message draft with a red pen
      The limit

      Polish does not find the right person

      Clean writing, wrong reader

      A message that scores 100 and lands on someone who never had the problem still gets ignored. Writing quality is the second half of outreach. The first half is who you write to.

      Start from a named complaint

      Crowd Scope finds the people publicly describing the problem your product solves, with a username, a post URL and a date. Use that post as your first sentence, then run the draft through this checker.

      Common questions about this rewriter

      Is this really an AI rewriter?
      No. Nothing here calls a language model. The rewrite applies fixed text rules in your browser: it removes pleasantry openers, swaps stiff phrases, drops hedging words and splits long sentences. We say so on purpose, because a rule list is easier to trust than a black box.
      Is the tool free, and do I need an account?
      It is free and needs no signup. Paste, read the score, copy the rewrite. There is no usage cap on the page itself.
      Does my message leave my browser?
      No. The scoring and rewriting run in JavaScript on this page. The only thing sent is an anonymous count that the widget was used, with no message text.
      Where do the 8 rules and the thresholds come from?
      They are our own editorial checklist for cold outreach, not a published benchmark. The 120-word and 25-word limits are working thresholds we chose for short messages. We have not measured reply rates against them, so treat the score as a draft checker, not a prediction.
      Why does a score of 100 not guarantee a reply?
      The score only checks the writing. It cannot see whether the reader has the problem you describe. A clean message sent to the wrong person still gets ignored.
      What does the rewrite not fix?
      Anything that needs judgment. It flags self-introductions, buzzwords, extra questions and over-long messages for you to edit by hand, because a fixed rule cannot invent a concrete fact about your reader.
      Can I use it on LinkedIn DMs and Reddit messages?
      Yes. The rules are about length, opener, hedging and question count, so they apply to any short cold message. Greeting and sign-off lines are left alone.
      How do I get a specific first line instead of a generic one?
      Start from a real post. Crowd Scope returns the person, the post URL and the date, so your first sentence can quote what they said. That is the line no rewriter can write for you.

      Write to people who already said they have the problem

      Crowd Scope returns named posts, dates and URLs from Reddit, X, LinkedIn and more, as a CSV in under 5 minutes.

      See how Crowd Scope works
      crowd·scope
      Start your investigation