Product · MCP server

Run DM Handled from Claude, or any AI agent

DM Handled is built to be operated by an AI agent, not just opened in a browser tab — read your inbox, reply, take over a conversation, manage its voice — the whole portal, open to your own AI.

What it's for

An MCP server lets an AI agent — Claude, or any other agent that speaks the Model Context Protocol — do what you'd otherwise do in the console: read what's waiting, reply, take over a conversation, review and approve voice examples, check what needs a human.

  • Built with the same account boundaries as the console — a key only ever sees one account's data
  • Read tools: account status, conversations, knowledge search
  • Write tools: reply, take a conversation over or hand it back to the agent, manage voice examples, see what's queued for a human
  • Configuration is deliberately excluded — an MCP key can't change billing, channels or settings
// illustrative — the shape of the calls, not a published API reference { "method": "list_conversations", "params": { "status": "needs_human" } }

Why this instead of just a dashboard

A dashboard is one more tab to remember to open. If the agent you already work in can read what's waiting on DM Handled and reply to it directly, the console becomes optional rather than mandatory — and an operations team already living inside Claude or another AI agent doesn't have to leave it to work.

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.