Run your email campaigns from an AI agent (MCP)
Most email tools assume a human is clicking through the dashboard. That breaks down the moment you want an AI agent to own a recurring send: models can't reliably drive a UI built for people, and screen-scraping automation is brittle. NewsletterLane takes a different approach — every operation you can do by hand is also exposed as an MCP tool, so an agent can run the whole workflow over a clean, typed interface.
What MCP gives an agent
The Model Context Protocol is a standard way for AI agents to call tools. NewsletterLane ships a remote MCP server whose tools are a thin wrapper over the same REST API your code would use. That means an agent gets the same capabilities, the same tenant scoping, and the same validation — there is no second, weaker code path. Connect an MCP-capable client (Claude, or anything that speaks the protocol), authenticate with a per-tenant API key, and the agent can list, create, and send.
Because the tools mirror the API one-to-one, an agent never has to guess at form fields or wait on a page to render. It calls create_list, gets back a list ID, and moves on. Every call is scoped to your organization, so an agent can only ever touch your own data.
The workflow end to end
A realistic agent run looks like this:
- Create a list for the audience you're targeting.
- Import subscribers from a CSV or an upstream source, with column auto-mapping and double opt-in respected.
- Draft a campaign from a brief — the agent writes the copy and assembles it with the block builder or raw HTML.
- Send from your verified domain, so DKIM, SPF, and DMARC alignment stays intact and one-click unsubscribe (RFC 8058) is included automatically.
- Read engagement — delivered, open, click, bounce, and unsubscribe counts, measured per unique recipient — and report back.
The agent doesn't need to know how the durable sending queue, rate limiting, or retries work. It hands off the campaign and NewsletterLane delivers it reliably. NewsletterLane sends broadcast campaigns, so an agent owns the whole send, not a triggered drip — which is exactly what you want when a model is composing and dispatching a newsletter on your behalf.
Where to start
Give your agent an API key, point its MCP client at NewsletterLane, and start with a small test list. See the features for the full tool surface, and the use cases for more patterns — from agent-run newsletters to backend automation. When you're ready to scale the sending volume, compare the plans on pricing.