Free tool · first DMs

What should your first automated DM say?

Tell it what started the chat and what you’re sending, and get openers written the way people text. Paste your own opener to check it for the tells that make a DM read like a robot wrote it.

What started the chat?
What’s the first DM for?
Reply-first delivery

Saw you commented GUIDE. Want the sourdough starter guide? Reply YES and I’ll send it over.

No robot tells found.

Either-or delivery

Saw you commented GUIDE. I’ve got two versions of the sourdough starter guide: reply 1 for the quick one or 2 for the full one.

No robot tells found.

Requests heads-up

Saw you commented GUIDE, and the sourdough starter guide is ready for you. Reply here so I can send it, since this chat may be in your message requests.

No robot tells found.

Paste a first DM you already use, or edit one from above. It’s checked for “a comment on a post or reel”.

  • FixIt opens with a stock phrase. Start with the answer. (“Thanks so much for reaching out”)
  • FixIt uses a word people don't text. Use the everyday word. (“comprehensive”)
  • FixIt has more than one exclamation mark. Keep one at most. (“2 exclamation marks”)
  • FixIt doesn’t ask for anything back. A comment earns one private reply, so end on something they can answer in a word: their reply is what lets you send the next message.
  • TipThe link is in the first DM. Asking for a one-word reply first gets more replies, opens the 24-hour window and moves the chat out of message requests; send the link after they answer, on a button.
How it works

Three steps, in your browser.

  1. 01

    Say what started it

    A comment earns one private reply, so its opener has to earn an answer. A DM or a story reply opens a 24-hour window, so its opener can do more.

  2. 02

    Pick the goal

    Send the thing, start a conversation, find out if they’re a fit, or get their email. Each wants a different first line.

  3. 03

    Check before you use it

    Every opener, written here or pasted in, is checked for stock phrases, words nobody texts, stacked exclamation marks, a name used twice, too many ideas, and a missing ask.

Methodology

How this tool gets its answer.

Openers are built from the shapes in our 12-openers guide and filled with your words; no AI model writes them. Every opener, written here or pasted in, is checked against HyperDM’s own list of the phrases that make a DM read as machine-written (stock openers and closers, customer-service phrases, words nobody texts, filler, stacked maybes, more than one exclamation mark or emoji, a dash used as punctuation, no contractions), plus a first name used more than once, more than three sentences, and an opener after a comment that doesn’t ask for a reply. The checks are about how a message reads. They can’t tell whether it’s accurate or right for your audience. The optional Write more with AI button is the one exception: it sends what you typed to OpenAI (gpt-4.1-mini) to write more openers, runs each one through the same checks on our server and leaves out any that raises a warning. We don’t store what you typed. It’s limited per visitor, and it switches off for the day once a fixed daily budget is spent.

Questions

Asked often.

Say who you are in context (“saw your comment”), what they get, and one small thing to reply. After a comment that last part matters most: you get one private reply, and their answer is what lets you send the next message.
After a comment, ask for a one-word reply first and send the link after they answer. A reply opens the 24-hour window and moves the chat out of message requests. Inside a conversation they started, the link can go in your first reply.
Write the way you text: short lines, contractions, one idea per message, a first name at most once, no “thanks for reaching out” and no stacked exclamation marks. Then read it out loud before you publish it.
Once, near the start, and only where the sentence still reads well if the name field fills in a handle instead of a name. More than once reads like a mail merge.
Yes, and you should. They’re starting points filled with your words. Paste your edited version back into the checker to see whether anything robotic crept in.
Not by default: the openers are built from fixed shapes in your browser. Write more with AI is optional. It sends your answers to OpenAI for more openers, and any opener the robot-phrase check flags is left out before you see it. We don’t store what you typed.

Got an opener people will answer? Send it on every keyword comment.

Build the comment-to-DM flow it starts, with typed replies understood as well as button taps.

Get HyperDMSTART WITH 14 DAYS OF PRO · THEN 50 CONVERSATIONS A MONTH FREE · NO CARD