Get API key
Email for AI agents

Give your AI agent
its own inbox

Need an email address for your AI agent? Agents can browse, click and fill in forms, then get stuck at “check your email”. BlinkInbox gives them real addresses they create themselves, and hands back the code or link as soon as the email lands.

Connect in one click Create a free account

MCP endpoint: https://api.blinkinbox.club/api/mcp

Capabilities

What AI agent email can do

Seven tools over the Model Context Protocol, and the same features over a REST API.

Create an address on demand

A real, working address in one call. Pick a readable prefix, add a label, use your own domain, and choose how long it lives (up to 30 days on Scale).

create_inbox

Wait for the email

The agent blocks until a matching message arrives, filtered by sender or subject. No polling loops, no guessing how long to sleep.

wait_for_email

Get the code or link, already parsed

OTP codes like 482913 or G-4F7K2Q and verification or magic links are pulled out of every message, so the agent never has to dig through HTML.

get_verification_code

Read the whole message

Subject, sender, text or HTML body, every link and attachment names. Enough to summarise a newsletter or check a receipt.

get_message

Clean up after itself

Delete an inbox when the task is done. Anything left over expires on its own, and mail is removed 24 hours after expiry.

delete_inbox

Hand off to your systems

The same inboxes are available over the REST API, and every new email can fire a signed webhook to your backend, n8n or Zapier.

REST + webhooks
How it works

Email for your AI agent in under a minute

1

Add the server

Paste the MCP URL into Claude, Claude Code, Cursor or VS Code. No API key to copy.

2

Click Allow

Sign in to BlinkInbox and choose full or read-only access. The app shows up in your dashboard.

3

Ask in plain English

“Sign up for the trial and give me the code.” The agent picks the right tools by itself.

Use cases

Agents that act for you

Your agent hits “check your email” and keeps going instead of handing the task back to you.

Sign up and verify

Create accounts for trials, betas and tools the agent is evaluating, then confirm them with the emailed code.

Finish browser tasks

Browser and computer-use agents complete flows that stop at an email step: downloads, waitlists, gated docs.

Magic-link sign-in

For apps with no password, the agent waits for the sign-in email and opens the link itself.

Subscribe and summarise

Give the agent a longer-lived address for newsletters, release notes or price alerts, and let it report what matters.

Use cases

Agents that test your product

Point a coding or QA agent at your app and let it go through your flows like a real new user.

Walk through onboarding

Sign up as a fresh user on staging, confirm the email and report every step that broke or confused it.

Review your emails

Catch broken links, unfilled {{placeholders}}, wrong product names or a missing unsubscribe link before customers do.

Check every email flow

Password reset, team invites, email change and 2FA codes, each with its own clean inbox.

Test as many users

A fresh address per run means no shared test account, no leftover state and no “email already in use”.

Use cases

Automations and pipelines

No chat window needed: the REST API and webhooks run the same flows from code.

End-to-end tests in CI

Playwright, Cypress or pytest create an inbox, trigger the email and long-poll for the code.

Email delivery monitoring

A scheduled job signs up every hour and alerts you if the verification email is slow or never arrives.

Email-triggered workflows

A webhook fires on every new email with the code already parsed, ready for n8n, Zapier or your own LLM step.

Seed test accounts

Fill staging with real, verified users for demos, load tests or support reproductions.

Try it

Prompts to start with

Copy one into your agent once BlinkInbox is connected.

“Create a BlinkInbox address, sign up for the free trial at acme.dev and tell me the verification code when it arrives.”
“Go through sign-up on staging.myapp.dev as a new user. Confirm the email and list anything that looked broken.”
“Request a password reset for the account using the inbox you just made, open the reset link, and check it works.”
“Make an inbox that lasts 7 days, subscribe it to the Acme changelog, and summarise each new email for me.”
“Read the welcome email our app just sent and check every link, the sender name and the unsubscribe footer.”
“Delete all the inboxes you created today.”
Guardrails

Built for agents you don’t fully trust yet

Receive-only

Agents can’t send, reply to or forward mail, so a BlinkInbox address can’t be used to spam anyone.

Read-only option

When you connect, you choose the access: read-only agents can read mail but can’t create or delete inboxes.

Untrusted content is labelled

Email bodies are marked as third-party data in every tool result, which helps agents resist prompt injection.

Dangerous files never stored

Executables, scripts, disk images, macro documents, HTML and SVG attachments are blocked before the agent can see them.

Revoke anytime

Every connected app is listed in your dashboard. Revoking it cuts off access immediately.

Nothing lingers

Each account sees only its own inboxes, and mail is deleted automatically after the inbox expires.

Use BlinkInbox for accounts you’re allowed to create, and follow the terms of the sites your agent signs up for. See our terms.

FAQ

Email for AI agents, answered

How do I set up an email for my AI agent?
Connect BlinkInbox as an MCP server (one URL, then click Allow), or call the REST API with a key. Your agent can then create its own addresses with create_inbox and read what arrives, including verification codes and magic links. The free Developer plan includes 3 inboxes in any 24 hours.
What is an email for AI agents?
It is a real email address that an AI agent can create and read by itself. BlinkInbox exposes inboxes as MCP tools and a REST API, so an agent can make an address, use it in a sign-up form, wait for the email and read the code or link without a person copying it over.
Which agents and apps does it work with?
Any client that supports remote MCP servers. Claude (web and desktop), Claude Code, Cursor and VS Code connect with one click through OAuth. Other clients can use an API key, and custom agents built on any framework can call the REST API.
Can my agent send email?
No. BlinkInbox is receive-only. Agents can create inboxes and read mail, but they can’t send, reply or forward.
How long does an agent’s inbox last?
The agent chooses when it creates the inbox, up to 60 minutes on the free plan, 7 days on Pro and 30 days on Scale. Mail is deleted 24 hours after the inbox expires.
Is it safe to let an agent read email?
Treat email as untrusted input. BlinkInbox labels message bodies as third-party data, blocks dangerous attachments and lets you grant read-only access. Keep human approval on for sensitive actions, and revoke apps from the dashboard at any time.
What is an MCP email server?
A Model Context Protocol server that exposes email as tools an AI agent can call. BlinkInbox’s MCP email server gives agents create_inbox, wait_for_email, get_verification_code and four more tools, over remote HTTP with OAuth sign-in.
Can my agent use my own domain?
Yes, on Pro and Scale. Verify a domain such as agents.yourco.com with one MX and one TXT record, and the agent can create any address on it.

Your agent’s inbox is one click away

Free to start, with 3 inboxes a day on the Developer plan.

Set up MCP REST API docsPricing