Account access and recovery
Open the console and sign in with your email and password. At signup, first select your intended use and accept the current beta terms. Cold outreach, prospecting, bulk marketing, inbox warmup and spam are prohibited; selecting these prevents account creation. Detected prohibited use will result in an account ban. A separate outbound product waitlist is available. Registration requires the operator and privacy contact to be configured. Existing users can continue connecting with an account API key. From Settings, a key-owned workspace can be linked once to a password login when registration is open.
Signup and password changes show a one-time recovery code. Save it in a password manager: email-based recovery is not yet available, and contact email addresses are unverified. Recovery resets the password, replaces the code and revokes earlier browser sessions. It does not revoke separately managed API keys.
Give each agent an identity and scoped key
Select New agent, choose a name and email domain, and save the returned key immediately. An agent is a mailbox identity for your own software, not an AI runtime hosted by Cyberpigeon. The Agents view lets you inspect its mail and pause or resume outbound sending. Pausing does not retract an email already being sent.
Use API keys to create account-wide credentials or keys limited to one agent with read, send, or both permissions. Key values are only shown on creation; the table shows the prefix, scope, revocation status and last-used time. Use a separate key per integration so it can be revoked independently.
Understand sent and received messages
Messages lists messages across the workspace. Filter by agent, direction and delivery status (including everything that needs attention), or search subjects and sender/recipient addresses. The seven-day chart counts messages created on each UTC day; the totals cover the account's stored history. Outbound counts include queued and failed mail, so they are not a delivered-email metric.
Select a subject to inspect the body, metadata, attempts and delivery timeline. Text is the default. The restricted HTML preview removes scripts, images, forms, links and styles to prevent remote tracking or active content; it is not a pixel-perfect reproduction. Very large messages are previewed from their beginning, and very deeply nested HTML is shown as text; a notice marks either case. Received attachments are separate provider downloads.
For API access, account keys can use GET /v1/messages with mailbox_id, direction, status, search, and cursor pagination. GET /v1/messages/{messageId} returns details and timeline. Agent keys use the mailbox-scoped endpoints in the API guide.
When a new account can send
Platform sending and receiving are enabled. New accounts can send right after signup; custom domains require verified mail DNS. Signup requires a permitted intended use: development with your own test recipients, support replies, or expected transactional messages. A mailbox on a custom domain sends once that domain’s mail DNS is verified. Automatic abuse checks can pause an account’s sending for review, returning 403 sending_review_required; contact support to challenge a hold. An accepted send request returns 202 for queued mail; confirm its actual outcome in message history.
Export, privacy requests and closure
In Settings, download a newline-delimited JSON export of account data, agents, messages and events. It excludes credential secrets. You can submit a privacy request and read operator responses in the same view.
Account closure requires your password and a typed confirmation. It revokes sessions and keys, pauses agents and starts live database erasure after a grace period and domain cleanup. Provider-held copies, recipients' copies and retained backups have separate handling. Read the privacy notice, beta terms and DPA preparation page before using personal data in the beta.