Product

One agent. Three inboxes. One set of rules it never breaks.

DM Handled reads every inbound message on Instagram, Facebook and WhatsApp, answers it in your voice inside Meta's own rules, and hands the conversation to you the moment a person is genuinely needed.

See it work

Want to see exactly how it'd respond?

Sign up free, connect your accounts, and watch it reply to your own recent DMs — in your voice — before a single message ever goes out. This shows the shape of a reply: it asks a real follow-up, remembers what was just said, and moves the conversation toward something useful instead of closing it down.

See how pre-flight replays your own real DMs →

AO
Amara OkaforReplied to your story
23:41:12 left

Where it answers

Four places. One inbox.

Instagram, Facebook and WhatsApp all work the same way underneath — reply inside Meta's 24-hour window, never start a conversation, hand over when a person is needed. The website widget puts the same agent, same knowledge and same voice on your own site.

Instagram

DMs, story replies and comment-started conversations, answered inside the 24-hour window.

How Instagram works →

Facebook

Messenger conversations on your Page, including the ones that start from a comment or a Click-to-Messenger ad.

How Facebook works →

WhatsApp

The same rules, plus Meta's own per-message pricing — billed to you directly, never marked up.

How WhatsApp works →

Website widget

The same agent, trained on the same knowledge, answering questions directly on your own site.

How the widget works →

What it learns from

Nothing it wasn't told

Two things shape every reply: how you sound, and what you actually know. Both come from you, both are visible in the console, and both can be changed at any time.

Voice

It learns your tone from real examples of how you write, so replies sound like you rather than like software.

How voice capture works →

Knowledge base

It reads your website, then you fill the gaps — exact prices, exceptions, policies. It never answers from anything else.

How the knowledge base works →

What keeps it safe

Compliance is checked before you trust it, and rebuilt to keep checking itself

Compliance & guardrails

Inbound only, official Meta APIs, no logged-in browser session, the EU AI Act disclosure that cannot be switched off.

The full account-safety page →

Pre-flight

Runs a safe dry test of the agent over your own recent conversations, so you can read what it would have said before it answers anyone for real.

More on pre-flight →

MCP server

Read your inbox, reply, and take over a conversation from Claude or any AI agent — the whole portal, open to your own AI.

More on the MCP server →

See it before you switch it on

Sign up free and read exactly what it would say before it answers anyone.

Free to set up.