Send a RelayLink briefing to yourself

One RelayLink account can hold Claude at work and ChatGPT at home. Address a package to your own email and it lands as unread inbound until any of your assistants pulls it — no second account, no contact request, no email.

3 min read

You worked something out with Claude at the office. Tonight you want ChatGPT at home to carry it forward โ€” same decision, same assumptions, same open questions โ€” without pasting a transcript, and without registering a second RelayLink account just so you can request yourself as a contact.

Address the package to your own email.

One account, as many assistants as you want

A RelayLink account is an email address, not an assistant. Connect Claude, issue a named key for an agent framework, connect ChatGPT on another device. Same inbox. Same contacts. Same threads.

The consent gate exists to keep strangers out of an inbox. You are not a stranger to yours. When the recipient is the sender, the gate lets the send through. Everything else is the ordinary path: draft, approve, confirm, same labels on who wrote what.

The landing page's "just you, across your AIs" cell is this use and nothing more.

What happens after you send

The thread has one participant. You cannot sit on it twice. A later draft on that thread still resolves to you.

It shows up as inbound, including to the assistant that sent it. Your inbox normally hides what you sent. A hand-off is the exception: it stays unread until any assistant on the account pulls it. Handed off, not yet picked up.

No email goes out. Notifications go to participants other than the sender, which here is nobody. No magic-link page, nothing in a mailbox. The portal inbox is how you pick it up without an assistant. confirm_send says so, rather than promising mail.

thread_status has a different answer. Ordinarily, if you were the last sender, the thread is awaiting them. On a hand-off you are always the last sender, so that rule would never clear. Instead: awaiting you until the latest package has been pulled, then idle.

How your assistant should treat it

Instructions tell it the address is valid, and that a pulled hand-off is your words to carry forward โ€” not somebody else's request to act on. A package from a contact is inbound correspondence. A hand-off is you talking to yourself across sessions. Still labeled. Still approved. Not a stranger's brief.

What this does not do

It does not bypass consent for other people. A new thread to someone else still needs an accepted pair if they have an account. A block still closes a thread that was already open.

It is also not a special tool or a new field. You draft to your own address the way you draft to anyone else.

Connect both assistants to the same account (setup), draft a package to your own email, approve it, and on the other side ask what is waiting. What's new and the inbox is the call that surfaces it.

Frequently asked questions

Do I need a second RelayLink account to move work between two assistants?
No. An account is an email address. Any number of assistants connect to it — a key each from your account page, or one OAuth grant each. Address a package to your own email and it goes through as a hand-off.
Does a self-addressed package send me an email?
No. There is no other participant, so no notification is sent and no magic-link page exists. Your portal inbox and your assistants' check_inbox are how you pick it up.
Will thread_status say I am still waiting on myself forever?
No. A hand-off is reported as awaiting you until the latest package has been pulled, then idle. The ordinary rule would have reported awaiting them forever, because you are always the last sender.