Managing a template library across email, SMS, push, and other channels involves more operational work than most people expect.
There's the actual content: writing, designing, reviewing. But there's also the surrounding workflow: approvals to sign off on, localizations to set up, drafts to audit, tags to organize, templates to preview before they go out. The result is a lot of context switching.
Dyspatch's new MCP server integration changes that. MCP, or Model Context Protocol, is an open standard that lets AI clients like Claude connect to external tools and data sources. Your AI client can now connect directly to your Dyspatch workspace and act on it, not just talk about it.
It can read your template library, manage drafts, handle localizations, and run approval workflows, all as part of a normal conversation. It can act on your behalf in a Dyspatch workspace.
You don't need to describe your templates to your AI client. It can see them. You don't need to copy and paste content back and forth. It can act on it directly.
That's a different kind of integration than what most teams are used to. And we’re just getting started.
The workflows that change
Here's what this looks like in practice across a few common scenarios.
Streamline campaign approvals
If your team uses Dyspatch's approval workflow, you know how it usually goes when you’re launching a big campaign. There are drafts sitting in PENDING_APPROVAL, you need to review each one, and approving them means opening each draft individually, checking the compiled content, and moving it through.
With the MCP integration, you can ask your AI client to pull all pending drafts, summarize what each one contains, and flag anything that looks off. Then approve the ones that are ready, in the same conversation. It compresses a workflow that usually spans multiple tabs and manual clicks into a single thread.
Setting up a new market
Localization work tends to be one of the more tedious parts of template operations. For every new language or region, you need to create localizations across multiple drafts, populate them with translated strings, and prepare them for handoff to your translation team.
With the MCP server, you can walk through that entire process conversationally. Create localizations for a new language across several drafts, seed the initial translation strings, and lock the drafts for translation, without leaving your AI client. It's the kind of setup work that used to mean a lot of repetitive clicking, done once in a thread.
Auditing your template library by asking a question
Ask your AI client to list all published templates across every channel type, summarize what each one does, and surface anything with a stale updatedAt date. What used to be a manual exercise — pulling exports, building a spreadsheet, cross-referencing dates — becomes a question you ask and get a useful answer to.
This is especially valuable when you're onboarding onto a new account, doing a pre-launch review, or trying to understand the full scope of what your team is actually sending.
Preview a template with real variable data before you send
The render_template tool lets you render any published template with real or sample variable data and see the compiled output directly in your AI client. If you want to do a quick sanity check on how a template will look with a specific set of variables before it goes out, you can do that without switching over to Dyspatch and setting up a preview manually.
Who this is for
The MCP integration is most useful for two roles: developers managing the communications infrastructure, and marketing ops folks handling approvals, localizations, and GTM preparedness.
If you're a developer or engineer who works on the tooling and infrastructure side of your team's communications stack, this gives you a way to manage Dyspatch operations directly from your AI client, and it's quick to set up.
If you're on the marketing ops or content side, the person who manages approvals, keeps localizations organized, and makes sure the right drafts are ready to go out on time, this is about reducing the operational overhead of that work. Less switching between tools, more time spent on the work that actually requires judgment.
In practice, both roles often touch the same workflows. The MCP integration sits in the middle.
Getting set up
Setup is straightforward. The Dyspatch MCP server runs via npx, so there's no installation step beyond having Node.js available. You point your AI client at the server, provide your Dyspatch API key, and you're connected.
Full setup instructions, including how to configure it for Claude Code, Claude Desktop, or any other MCP-compatible client, are in the Dyspatch MCP documentation.
The bigger shift
AI tools are most useful when the boundary between thinking about work and doing it starts to disappear. That's easy to say, but most AI integrations don't actually go beyond suggestions, summaries, generated content, which leaves you having to then go execute somewhere else.
The Dyspatch MCP integration is different because it connects the thinking to the doing. Your template library isn't something you describe to your AI client anymore. It's something your AI client can see, manage, and act on, in the same place where the planning is already happening.
If you want to see what that looks like in your own workflow, get a demo or head to the MCP documentation to get started.
