Blog
Blog
Dated writing from AgentiSend: how the send guardrails work, why they are built that way, and where they stop. 6 posts.
Categories: Agents, Guardrails, Receiving
- An agent that signs itself up
2026-10-21
An AI agent signs itself up for AgentiSend with POST /agent/sign-up. The key emails only that human until POST /agent/verify.
- Mail an agent reads is where it gets hurt
2026-10-21
Mail an agent reads on AgentiSend carries inbound_verdict. The label is clear, review, or unavailable, and the body stays data.
- One MX record, and your agent can receive
2026-10-21
One MX record on a verified AgentiSend domain is what lets an agent receive mail there. Each message arrives as email.received with the reply already separated.
- How loop detection decides a send is a repeat, and what it holds
2026-09-29
How the loop guard decides a send is a repeat, what it holds for a person, and the send it still delivers.
- What an agent should do with a refusal: code, fix, retry_after_seconds
2026-09-29
A refusal carries a code, a message and a fix. Wait only when retry_after_seconds is present. Three decisions stay with a person.
- Why every API key is created with a budget
2026-09-28
Every new AgentiSend API key starts with a send budget equal to the plan inclusion. The key cannot raise it, and sending stops when the budget is used.