Can I drive Broadcast from my own app or an agent?

Yes, three ways, and they all sit on the same permissions model.

REST API

Full coverage of the things you would automate: subscribers, broadcasts, segments, sequences, templates, opt-in forms, email servers, webhook endpoints, and transactional email.

Start with API Authentication for tokens, scopes, and the rate limit. Read API Response Warnings too: write requests can succeed while doing something other than what you asked, and the warnings array is where that shows up.

Creating an access token with per-resource read and write permissions for broadcasts, subscribers, sequences and more
Tokens are scoped per resource, so an integration gets only the access it actually needs.

Command line

The CLI wraps the same API for scripting and one-off operations without writing HTTP calls.

MCP server

An MCP server exposes Broadcast natively to Claude Code, so an agent can work against your install directly. The agents overview compares the three approaches, and setup links are time-limited credentials for onboarding an agent without pasting a long-lived token.

The Webhook Endpoints page listing endpoints and the event types each one subscribes to
Outgoing endpoints, each subscribed to the events your own application cares about.

Events going the other way

For your app to react to what happens in Broadcast, use outgoing webhooks. They cover delivery, opens, clicks, bounces, and complaints, with signature verification and a documented retry schedule.


If an endpoint you need is missing, tell us which one.