What if an AI agent was your teammate? With the new dotCMS MCP Server, you can work side-by-side with AI agents and assistants to draft, edit, manage, or audit content, and always with your rules, your workflows, and your brand voice.
What Is an MCP Server?
An MCP, short for Model Context Protocol, is a secure connector that lets AI agents interact with any system the same way a human user would.
It gives your AI tools the context they need to understand:
What data exists
How that data is structured
What actions they can perform
What rules and permissions they must follow
Instead of “guessing,” your AI assistants have a standard way to read, write, and manage data, safely and with granular control and oversight.
In dotCMS, the MCP Server does exactly that for your content teams. It automatically brings in your context, then lets trusted AI agents:
Discover your sites, content types, and workflows
Draft and update content using your brand rules
Run workflow actions with full oversight
Your Secret to AI That Works: Context First
AI only works it best by understanding the world around it.
That’s why the dotCMS MCP Server doesn’t just connect your CMS, it shares your sites, content types, and workflows automatically.
This context keeps your AI grounded in reality, so its drafts follow your structure and workflows instead of guessing.
How to Set Up the MCP Server
Here’s how to get started.
Get a dotCMS environment Cloud, on-prem, or our demo site — they all work.
Create an AI User and API Token
In your dotCMS admin, create a user just for your AI agent.
Give it only the permissions it needs.
Generate a secure API token.
Configure the MCP Server in your AI tool (like Claude Desktop).
Open Claude Desktop
Go to Settings > Developer
Click on Edit Config
This will open the folder with the file claude_desktop_config.json
Edit that file and add the following configuration:
{
"mcpServers": {
"mcp1": {},
"mcp2": {},
"dotcms": {
"command": "npx",
"args": ["-y", "@dotcms/mcp-server"],
"env": {
"DOTCMS_URL": "https://your-dotcms-instance.com",
"AUTH_TOKEN": "your-auth-token"
}
}
}
}Restart Claude Desktop and start prompting about dotCMS.
How to Use It: Make AI Work for Your Whole Team
Connecting your MCP Server is just the first step. Now it’s about making AI part of your team’s daily routine in a way that keeps your content on-brand, consistent, and compliant.
1. Pick Real Tasks
Start with everyday jobs that slow you down:
Draft a blog post from bullet points.
Rewrite outdated product pages.
Search for content with missing required disclaimers.
Generate meta descriptions for old content.
Keep it simple. Test with one piece at a time. Small wins build trust.
2. Build a Shared Prompt Library
Your AI only writes well if it knows your brand voice, tone, and rules. One-off instructions won’t cut it. Create a shared prompt library your whole team can use.
What goes in it:
Voice & Tone: How you sound. Friendly but professional? Formal? No slang?
Style Rules: Short sentences, active voice, words to avoid.
Compliance: Legal disclaimers, terms never to edit, regional requirements (like GDPR or HIPAA).
Approved Words: Product names, trademarked phrases.
Example Prompts:
“Write in a clear, neutral tone for finance customers. Keep it factual. Add our standard compliance note at the end.”
“Friendly and conversational for our blog. Always use approved product names exactly as listed.”
“Rewrite this page for EU readers. Formal tone. Include GDPR statement. Never remove legal footers.”
Keep your prompt library in an easy-to-find doc or team folder. Update it as you learn what works.
3. Test Small, Review Together
Try small tasks first. Use dotCMS workflows. Keep a human review step. Would you publish a draft from a new writer without edits?, I’m pretty sure the answer is NO, so don’t do it with AI either.
4. Share and Improve
Good teams share what works.
Show your team which prompts get the best results.
Refine them over time.
Make prompt writing part of onboarding for new team members.
The better your prompts, the better your AI will perform — and the less rewriting you’ll need.
Result: Your AI does the busywork. You focus on strategy and final edits. And every word stays true to your brand and industry rules.
Examples You Can Try Right Now
Need ideas? Here are a few ways to use your AI assistant:
Drafting First Versions
Turn outlines into first drafts.
Expand bullet points into paragraphs.
Rewrite a long post for a new audience.
Bulk Updates
Update old meta descriptions to match your new SEO strategy.
Find and fix product names that changed last quarter.
Refresh headlines for better consistency.
Repurposing Content
Split a long article into social posts.
Create newsletter snippets from your latest blog.
Rewrite a landing page for a new region.
Compliance Checks
Search for content missing required disclaimers.
Flag old content that mentions outdated terms.
Double-check tone and style for brand fit.
Build Good Habits With Your Team
AI works best when your team works together.
Test tasks in small batches first.
Always review AI drafts before publishing.
Use your workflows. Keep your human-in-the-loop step.
Share your best prompts so everyone benefits.
When everyone writes better prompts and shares what works, you build a habit. The more you do it, the faster you get.
Stay in Control
The dotCMS MCP Server keeps your governance rock solid with dotCMS content history and logs. Every AI action is:
Logged: You see who did what.
Traceable: No mystery edits.
Reversible: Roll back anytime.
Permission-based: Your AI works only where you allow it.
Your workflows don’t change — they just get faster.
Work Smarter, Stay Compliant
dotCMS is built for teams that need to publish faster without sacrificing governance. The MCP Server makes your AI work inside the rules you trust, so you spend less time on grunt work and more time on ideas that move your brand forward.
Ready to get started? Learn how to set up your MCP Server and put your new AI teammate to work.