TwinsGmail
A Gmail twin for testing AI agents.
Test agents that read, triage, draft and send email. Every mailbox is private and seeded, so nothing reaches a real inbox.
A stateful copy of Gmail, built for testing agents.
It accepts the same API and MCP calls, checks sign-in and input, and follows the same rules, but runs on its own data, so nothing reaches your real account.
- Messages, threads and drafts
- Send, reply and forward
- Labels, archive and search
- Attachments
- OAuth scopes and permission errors
Every run starts from a company that already has history. We seed:
- A mailbox for every person in the company
- Threads with real history between them
- Labels, unread state and starred mail
- Attachments and calendar invites
Use Gmail in staging, without using production.
Validate every action
The twin checks what the agent sends the way Gmail would: required fields, permissions, rate limits, malformed input. Mistakes fail here, not in your account.
Start from a realistic state
Describe the company and the situation in plain English. Drizz seeds the twin to match, and resets it to that state before every run.
Replay tasks across tools
The Gmail twin shares its people, customers and history with the other twins, so a task that touches three tools runs the way it would at work.
Swap the address. Your agent code stays the same.
Replace the live Gmail endpoint with the twin’s. Everything your agent already does keeps working, on data that is yours to break.
{
"mcpServers": {
"gmail": {
"type": "http",
"url": "<twin MCP endpoint>",
"headers": { "Authorization": "Bearer <token>" }
}
}
}REST works on the same data. Both are reset with one call.
Connected to the rest of the company.
The people in Gmail are the same people in Calendar, Slack and the CRM. A thread about a renewal links to the deal, the meeting where it was discussed and the Slack channel where it was escalated.
Read the guide: How to test AI agents before production →