agents can write code but they can't send an email so i gave them email ☑︎ the agent claims its own address ☑︎ you approve it in one click ☑︎ it sends, receives and replies over mcp or api claim your agents email at

yeah but we got grokbot emails now

yeah but what if you wanna use it for a non grok bit

fair enough, personally havent encountered any moments where agents need emails except for bots now but a cool side project nonetheless!

codex + computer use with gmail is what I have been using like a noob

Seems nice

Cool shi. Eagerly waiting for some more things: -opencode browser extension you showcased few days ago -opencode sign in (like chatgpt sign in). With access to both zen and go models

OpenCode Browse coming soon, idk opencode sign in isn’t coming soon

Appreciate the reply! Really hoping OpenCode sign-in drops soon—it could drive serious traction for Go plans with non-coders, especially alongside the $5 tier Dax mentioned. Quick question on BotMail: does it support aliases?

Nice!

Looks awesome! The GitHub flow doesn’t work for me - you sign in to GitHub and it just sends you to GitHub not back to sign up

Agent is checking

Thanks! On iOS safari

how do you keep a freshly claimed address out of spam folders on day one?

there is an automated classifier that runs on every email to make sure fraud and spam emails are not let through this is a pretty warm domain as well

a warm shared domain does a lot of the lifting there, good call

dkim alignment gets tricky when the agent sends but your app signs the key. a fresh address has no reputation so bounces tank deliverability fast even on a warm parent domain. do you keep a reputation pool per agent and rotate, or rely on the classifier?

grok @bot email?

Inbound reskin?

nope, inbound still exists but likely the two will get merged

Nice! How’d you make the product video?

agents can write code but stall on email. a claimable address with one-click approve is the staffing move that stuck for me

Let me know when you will propose the gmail sign-up 🙏

coming today doing a slow rollout

"agents can write code but they can't send an email" Such an incorrect statement it's incredible Agents can't natively send email is the correct statement. My agent has had it's own inbox with purelymail's api for the past 3 months.

Does the one-click approval also gate each outbound send?

The one-click approval is the right idea. Does the human see a draft before every send, or only the first time?

Up to you

Nice, so it's a setting. I'd keep drafts on for anything that goes to a client, and let it send on its own only inside my inbox.

Yeah just shipped drafts