Instagram DM automation: your rules, AI drafts

Fig. 1rules · @bistro.demodemo

Keyword “guide”
Comment → public reply + link in Direct
412 sent
No reply in 15 min
The safe draft goes, with a Stop button
23 sent
Heart on story reactions
A heart back on a reaction, once a day per person
61 sent
Keyword “menu” · next post
Ready before the reel goes out
waiting

@ana.lGUIDE 🙌sent to direct
↳ @bistro.demoSent it to your Direct 👋
@ben_kthis is so usefulno keyword
@cris.vguide pleasesent to direct
↳ @bistro.demoCheck your DMs!
@dmitri🔥🔥🔥no keyword
@eva.swhere's the guide?sent to direct
↳ @bistro.demoOn its way to your Direct

Comments 0

With “guide” 0

Sent to Direct 0

Waiting for pace 0

Pace 0/h

Refusals slow it down; a spam refusal pauses the account’s rules.

What runs on rules

Keywords in comments, Direct and story replies: your texts, no model. See Comment to DM.
A timer: if nobody answers in time, a draft the checks found safe goes out, after a warning with a Stop button.
Outside working hours: a template reply in Kazakh and Russian, and a draft for the morning.
A heart on a reaction to your story.

What the assistant drafts

A reply to every other Direct message, written from the whole conversation in your brand’s voice. Prices, complaints and promises go to a person. Comments get drafts too; they wait in the workspace, with one Telegram card per post.

Where your team works

In Telegram: one card per message with ✅ Send · ✏️ Edit · Don’t reply. Today the bot speaks Russian; its buttons are translated on this page. Instagram in Telegram.

Instagram’s rules, respected

Meta’s official login, no password. The 24-hour window and one private reply per comment are kept, and the pace slows when Instagram refuses. Apply for the beta.

FAQ

Questions

Can it answer everything automatically?

No. It sends on its own only what your rules allow. Every other message gets a draft and waits for Send, Edit or Don’t reply.

Can I stop a timed reply?

Yes. Before the draft goes, Telegram shows a warning with a Stop button.

Will replies say they are from AI?

By default replies carry no AI signature, and a sincere “are you a bot?” goes to the account owner.