REST API vs MCP: two ways to automate email
NewsletterLane gives you two programmatic ways to run email: a versioned REST API and a set of MCP tools. They expose the same capabilities — lists, subscribers, campaigns, sends, and analytics — so the choice isn't about what you can do, it's about who (or what) is doing it. This post explains when each surface fits.
The REST API: built for your code
The REST API lives at /api/v1 and ships with a full OpenAPI document, so you can generate a typed client in any language. Authentication is per-tenant: you issue an API key, scope it to your organization, and every request stays inside your own data. You can create and list resources, import subscribers, build campaigns, and trigger sends — the same operations the dashboard performs.
Reach for the API when a deterministic system is in control: a backend job that imports new signups nightly, a webhook handler that adds a contact when someone converts, or a CI pipeline that ships a release-notes broadcast. You know the exact sequence of calls, you want predictable error handling, and you want to pin behavior in tests. Because there's an OpenAPI spec, your client stays honest as the API evolves.
MCP tools: built for AI agents
MCP tools wrap that same REST API in a model-friendly way. Instead of reading OpenAPI and writing a client, an AI agent discovers the available tools at runtime and calls them directly. The tools carry descriptions and typed inputs the model can reason about, so an agent can decide which operation to call and when — without you hand-coding the flow.
Reach for MCP when judgment is part of the job: an agent that drafts a newsletter from a brief, decides which list to target, sends it, and summarizes the results. The agent isn't following a fixed script; it's composing the workflow. Under the hood it's still the same tenant-scoped operations, so you get the same safety guarantees as the API.
They're the same engine
This is the important part: MCP tools are a thin client over the REST API. There's one implementation of "create a campaign," one validation path, one permission model. Whether the caller is your backend or an autonomous agent, the behavior is identical. That means you can prototype with an agent over MCP, then harden the parts you want deterministic into API calls — no rewrite, no divergence.
Which should you use?
Use the REST API for fixed, repeatable automation you control end to end. Use MCP when you want an AI agent to make decisions and run the workflow itself. Most teams use both: scripted plumbing on the API, creative or judgment-heavy work on MCP. Both are available on every plan.
Explore the full surface on the features page, see real patterns in use cases, and check pricing when you're ready to send at volume.