Give your AI agent a real Mailfence inbox. Read, search, send, reply, schedule and organize mail over IMAP, on any MCP client, with the plan requirement and the two-factor caveat stated up front instead of discovered later.
Claude cannot reach Mailfence on its own. mcpemails is a hosted MCP server that connects it over IMAP at imap.mailfence.com on port 993, and SMTP at smtp.mailfence.com on port 465, so Claude, Cursor or any MCP client can read, search, send and organize that mail. Two conditions decide whether it will work: IMAP, POP and SMTP are available on Mailfence's paid plans only, not on the free account, and the login must be your primary address or username rather than an alias. Authentication is your Mailfence password, or better, a service-specific password generated under Settings, Account, Security. One inbox is free at our end, forever.
Mailfence serves IMAP at imap.mailfence.com on port 993 and SMTP at smtp.mailfence.com on port 465, both with SSL required, and POP3 on 995. Your own values are printed inside the account under Settings, then Account, then My connection settings. The part that decides everything else is your plan: protocol access is a feature of the paid tiers, and a free Mailfence account is webmail only, so no client, ours included, can log in to one. Authentication uses your Mailfence password by default. Mailfence also lets you generate a service-specific password per protocol under Settings, Account, Security, Services access, and switch off the protocols you do not use. The same screen can disable IMAP entirely, which is worth knowing in both directions.
Copy these into the connect form. mcpemails prefills them where it can, and retries the other standard transport if the first does not answer.
| Protocol | Server | Port | Security |
|---|---|---|---|
| IMAP (incoming) | imap.mailfence.com | 993 | TLS (implicit) |
| SMTP (outgoing) | smtp.mailfence.com | 465 | TLS (implicit) |
Verified 2026-08-31: imap.mailfence.com answered a TLS connection and identified itself as an IMAP server under that name. It advertises PLAIN authentication.
Sign in with: Your primary Mailfence address or your Mailfence username. Alias addresses are refused.
Nine action-based MCP tools: send, reply, forward, schedule, and organize. Your agent finishes the job inside Mailfence.
List, read and search across folders to find that invoice, summarize a thread, or pull the latest from a sender.
Compose and send real messages, reply in-thread, and forward, directly from your mailbox, not as a draft you finish by hand.
Queue a message to go out at the right time, so your agent can draft now and send on schedule.
Keep the inbox tidy: file mail into folders, flag what matters, archive the rest, or delete on request.
From sign-up to first AI email in a couple of minutes.
Sign up at mcpemails.com. The free plan connects one inbox, forever, with no card. You get an MCP URL and a key at the end of it.
Open Settings, then Account, then My connection settings. If your plan carries IMAP, POP and SMTP access, that page prints the servers and ports for your account. On a free account the access is simply not there, and no combination of hostname, port and password will change that. Mailfence documents protocol access as a paid-plan feature.
Under Settings, Account, Security, Services access, create a password scoped to IMAP and one for SMTP rather than reusing your account password, and disable POP, xDAV and any other protocol you are not using. Mailfence's own hardening advice is to do exactly this. Then add the mailbox in mcpemails with your primary address or username and that password.
Drop https://mcpemails.com/api/mcp into Claude, Cursor or ChatGPT, authorize, and your agent has the inbox.
Comparing providers first? See the email provider compatibility matrix
Mailfence gates IMAP, POP and SMTP behind its paid subscriptions. A free account is a webmail account. Because the server still answers on 993 and still refuses your credentials, the rejection is indistinguishable from a typo, and people spend an afternoon resetting passwords that were correct the whole time. Check My connection settings first: if it does not offer you protocol details, the plan is the reason and upgrading is the only fix.
Mailfence states this plainly in its own 2FA documentation: two-factor authentication does not apply to non-web connections such as SMTP, IMAP, POP, EAS, xDAV and XMPP. Turning it on protects your webmail login and changes nothing about a mail client, which keeps working with the password it already has. So your client will not break when you enable 2FA, unlike at most providers, and 2FA is not the control that limits a connected client. A service-specific password under Services access is, because it can be scoped to one protocol and revoked without touching your account password.
Mailfence's client documentation is explicit that only the primary email address or the username can be used for authentication. Aliases, extra addresses and mail on a second domain you have attached all deliver into the same mailbox, but none of them will get you through the login. Use the address the account was created with, or the bare Mailfence username, and put the address you actually want to send from in the identity rather than in the credential.
Unusually for this corner of the market, Mailfence's IMAP server advertises MOVE, UIDPLUS, IDLE and SASL-IR, so filing mail is a single atomic command and bulk reorganising is quick rather than a copy-flag-expunge dance. What it does not advertise is SPECIAL-USE, so nothing on the wire identifies which folder is Sent, Drafts, Trash or Junk and roles are inferred from names. It also offers only AUTH=PLAIN on IMAP, with no AUTH=LOGIN, while SMTP accepts both, so tooling hard-wired to LOGIN can read nothing and yet send fine.
Not directly. Claude has no built-in Mailfence connector. Connecting an MCP server such as mcpemails to a paid Mailfence account over IMAP gives Claude read, search, send, reply, schedule and organize access to it, and the same connection works in Cursor and other MCP clients without repeating the setup.
IMAP is imap.mailfence.com on port 993 with SSL, SMTP is smtp.mailfence.com on port 465 with SSL, and POP3 is on 995. The username is your primary Mailfence address or username, never an alias. Your own account prints the same values under Settings, Account, My connection settings. Both hosts here were verified from a live TLS connection.
No. Mailfence makes IMAP, POP and SMTP access a paid-plan feature, so a free account is restricted to webmail and to Mailfence's own apps. No third-party client and no MCP server can log in to a free account. Upgrading to any of the paid tiers turns protocol access on for that mailbox.
No, and Mailfence says so itself. Its documentation states that 2FA does not apply to non-web connections including IMAP, POP and SMTP. Enabling it secures the web login only. If you want to control what a connected client can reach, generate a service-specific password under Settings, Account, Security, Services access, and disable the protocols you do not use.
No. Mailfence's encryption relies on a key that your AI client does not have and that mcpemails never receives, so an encrypted message arrives over IMAP as an armoured ciphertext block. The agent can see the sender, subject and date and can move or flag the message, but the body is unreadable outside the Mailfence interface.
No. Messages are fetched live from imap.mailfence.com for each request and passed to your AI client. Credentials are encrypted at rest and message bodies are not retained. The AI client itself does receive whatever it is asked to read, which is the part to weigh before connecting a mailbox you chose Mailfence for in the first place.
Connect a paid Mailfence mailbox in minutes. One inbox free forever, no card required.