AgentMail Changelog

Inboxes can now carry a signature that AgentMail adds to every message they send. Your agents no longer have to remember to append contact details or branding, and every message from the inbox looks the same whether it was sent through the API, the MCP server, or SMTP.

Inboxes can now pause themselves at a time you choose. Set pause_at when you create an inbox, or later, and AgentMail pauses it when that time comes, with no scheduler on your side. This is built for inboxes sold or provisioned for a fixed term.

Calendar invitations your agent builds itself and sends from an inbox now go on that inbox’s calendar. Attendees’ replies are tracked on the event, the same as for invitations sent with the Calendar API.

message.received for an email with a calendar invitation now arrives with calendar_event_id already set, so your agent can answer the invitation straight from the webhook instead of fetching the message again.

Calendar is now available to every organization in US production, with no opt-in. Every inbox already has a calendar: your agents can create events, send and answer invitations, and react to start and end webhooks without joining AgentMail Labs first.

Agent Armor can now hold mail. In Hold mode, your agents hear about a message only once its scan finishes, and a message marked review stays held from them until someone releases it.

Agent Armor is now in AgentMail Labs. An organization admin can turn it on from the console, and every message your inboxes receive is scanned for attempts to manipulate an AI agent.

Agents can now check and join AgentMail Labs through the API. Any organization API key can check whether the organization is in Labs, and one with no permissions restriction can join it once the organization has decided to, without opening the console.

Calendar is now in AgentMail Labs, so your organization can turn it on itself instead of asking us for beta access. An organization admin joins Labs once in the console, and every inbox’s calendar becomes available to your agents right away.

Invitations are easier to keep up with. Every event the inbox can respond to now says how the inbox has answered, so an agent can find the invitations still waiting for a reply from the agenda alone. And an invitation email now records the calendar event it was applied to, so an agent can go from a message.received email straight to its event.

Apps now say what kind of app they are. Each app in the catalog can carry up to three categories, and List Apps takes a category filter, so an agent looking for a web search API can list only search apps.

Apps in the catalog now have a slug, a short name you can use in place of app_id. An agent can connect an inbox to Firecrawl with POST /v0/apps/firecrawl/connect, without listing apps to find its ID first.

Every inbox now has a calendar (private beta). Agents can schedule one-off and recurring events, send and receive standard calendar invitations by email, and get calendar.event.starting and calendar.event.ending webhooks as each event begins and ends, without running their own scheduler.

Humans can now claim an inbox their agent created without a human email. The human pastes the agent’s API key into the AgentMail Console, and the agent’s inbox becomes an organization they own, with sending unlocked and the agent’s key still working. Agents no longer have to wait for their human’s email address before they can send.

The API reference now documents pausing an inbox. The API already supported it: set an inbox’s status to paused to stop it sending and receiving mail without deleting it, and set it back to active to resume.

The two API key permissions behind AgentID sign-in now use the app noun: provider_connect is app_connect, and provider_share_owner is app_share_owner. Existing keys keep their grants. The old names are removed: requests that send them get a 400, and responses return only the new names.

Providers are now apps. Find an app, connect an inbox to it, and list the accounts your agents hold there through /v0/apps, client.apps, and agentmail apps. The provider names are removed, not deprecated: /v0/providers, client.providers, agentmail providers, and the provider_id and provider_name account fields are gone.

Ask Claude, ChatGPT, Cursor, Claude Code, or Codex to “create an account at Firecrawl for my agent” and it does, using an AgentMail inbox as the agent’s identity through AgentID. A new guide walks through creating accounts, getting API keys, and checking where your agent has accounts, and the AgentMail plugin ships a skill for it. A provider that is registered but not listed in the catalog can also be read and connected by its ID.

The API reference now documents expires_at on bearer API key creation. The API already accepted it: give an agent a key that stops working on a set date, so short-lived work never leaves a standing credential behind.

Agents can now sign up without a human’s email address. They get an inbox that receives email right away, and they attach a human later when they have one. Sending stays locked until a human is attached, so an agent can start receiving mail without waiting on anyone.