For AI agents

A helper that knows your email.

Connect the AI agent you already use and ask for help in plain words. It works in one workspace, with only the access you approve, and you can disconnect it anytime.

Connect

Connect in a minute.

getread’s MCP server lives at https://mcp.getread.com/mcp. Add it to any MCP client that supports remote servers with OAuth, and you’ll be asked to sign in.

  • Sign in with your password, Google or GitHub
  • Pick the one workspace it can work in
  • Approve the permissions it asks for
{
  "mcpServers": {
    "getread": {
      "url": "https://mcp.getread.com/mcp"
    }
  }
}
Most clients add a remote server with an entry like this. The exact format depends on your client.

You’re in charge

Before anything happens, you see exactly what your agent is asking for. Choose the workspace it can work in, then approve or deny.

It can draft and update campaigns, but it can’t send one. That part always stays with you.

The consent screen asks you to authorize Studio Assistant, lists the permissions it requests, lets you pick one workspace and offers Deny and Approve buttons.

Always visible

See what’s connected, and let go anytime.

Every connected agent shows up in Settings → Connected apps, with its workspace and permissions. Everything it does is attributed to that connection, so you always know who did what.

When you’re done, revoke its access with one click.

The Connected apps settings page shows Studio Assistant connected to the Acme Studio workspace, with its granted permissions, when it connected, and a Revoke access button.

Ask in plain words

What it can help with.

Ask the way you’d ask a teammate. It works with the same email you do.

  • CampaignsDraft and update campaigns, ready for you to review and send.
  • AutomationsBuild and adjust automations, like a welcome series.
  • Contacts and listsAdd, update and tidy up the people on your lists.
  • Templates and formsWork with your saved templates and signup forms.
  • AnalyticsLook up how your email is doing.
  • Transactional emailSend an order update or a receipt, once an owner or admin allows it.

Same rules for everyone

Helpful, and careful too.

Your agent works within the same limits as everything else in getread.

  • Owner or admin approval to sendOnly an owner or admin can let an agent send transactional email.
  • No accidental duplicatesEvery send from an agent carries an idempotency key, so a retry never sends twice.
  • The same limits as the APIAgent sends follow the same rate limits, suppression and sending policy as the API.
  • Unsubscribes stay unsubscribedPeople who unsubscribe, bounce or complain are suppressed, whoever is asking.

Questions about AI agents

Which AI agents work with getread?
Any MCP client that supports remote servers with OAuth. Add https://mcp.getread.com/mcp as a server, sign in, and approve the connection.
Can my agent send a campaign?
No. It can draft and update campaigns, but sending one is always up to you.
Can it send transactional email?
Yes, once an owner or admin allows it. Each send needs an idempotency key and follows the same rate limits, suppression and sending policy as the API.
Can it see all my workspaces?
No. You choose one workspace when you connect. To help in another one, connect again and choose that workspace.
How do I disconnect it?
Go to Settings → Connected apps and revoke its access. That stops any future calls. It doesn’t undo changes it already made, or email that was already accepted for delivery.
Does it cost extra?
No. Help from your AI agent is included on every plan, and any email it sends comes from your plan’s monthly pool like everything else. See pricing.

Bring your agent along.

Connect it in a minute, with only the access you choose.

[TRIAL TERMS]