Connect an agent.
Receive the signal.

1. Create your account and key

Open the console and sign in with your notification email — a 6-digit code is sent to that address. Then create an API key on the Keys page. The raw key is shown exactly once, so store it right away.

Each key is bound to that one email address: every notification goes there, and an agent can never override the recipient.

The console also shows your usage, recent events, and decision history at a glance.

2. Connect your agent with one prompt

On the console's Keys page, open the integration guide and copy the setup prompt. Paste it into your agent (CodeBuddy, Codex, Claude Code, WorkBuddy, or any MCP client) — the agent installs the local adapter, registers itself, and verifies the connection automatically.

From then on, the agent reports task completions, failures, blockers, and decision requests on its own, following the rules in its skill file.

Prefer wiring it yourself? The same REST API and stdio MCP adapter are available for any client — ask your agent to read the service's skill file for the exact endpoints.

If your platform can't poll, configure a signed webhook in the console (Settings): SIEmail pushes the decision to your URL the moment it is made or expires.

3. What lands in your inbox

• Task done — summary, verification, and next steps, not just a “finished” ping.
• Blocked or failed — what went wrong and what the agent needs to continue.
• Decision needed — options, trade-offs, and a recommendation, with a one-click choice in the email.

Duplicate reports are deduplicated by event ID, so a retrying agent never spams your inbox.

4. Your data and security

• API keys are stored as SHA-256 hashes; the raw key is shown exactly once at creation.
• A key is bound to one verified email; requests cannot override the recipient or sender.
• Event payloads are kept for 90 days and audit logs for 365 days, then purged automatically.
• Don't send passwords, access tokens, or customer data in events — keep summaries short.