Model Context Protocol
Open Standard

Your AI.
Inside your CMS.

The Model Context Protocol is an open standard that lets AI assistants connect directly to external systems. Instead of copying and pasting, your AI reads, writes, and acts. Axolot is the first CMS with native MCP servers.

Think of it as giving your AI a set of hands. It can navigate your CMS, modify content, query your database, and deploy changes — all from a single conversation in Cursor or Claude Desktop.

3

MCP Servers available

Free

Forever open source

<50ms

Response latency

Operations per session

The Workflow

From conversation
to live change.

01

You ask your AI

In plain English, inside Cursor, Claude Desktop, or any MCP-compatible tool.

"Update the hero title to reflect our new brand positioning."
02

AI calls the MCP

The AI identifies the right server and calls the relevant tool with structured parameters.

master.updateSlot(key: "hero.title", value: "...")
03

Server acts on Axolot

The MCP server authenticates and executes the operation against the Axolot API or database.

PATCH /api/v1/slots/hero.title → 200 OK
04

Change goes live

Your site reflects the update instantly. The AI reports back with a confirmation.

"Done. The hero title has been updated."

Real Example

A full site update.
One conversation.

What used to take 45 minutes across 3 different dashboards now happens in a single chat session. The AI has full context — your site structure, your content, your data — and acts with precision.

Cursor · AI Chat
List all pages on my site and find any hero titles not updated this month.
⚙ Calling master.listPages()...
⚙ Calling master.getSlots(filter: "hero.*")...
Found 4 pages. The /about and /services hero titles were last modified 47 days ago. Want me to update them?
Yes, make them more benefit-focused.
⚙ Calling master.updateSlot(about.hero.title)
⚙ Calling master.updateSlot(services.hero.title)
✓ 2 slots updated successfully
Server 01 of 03
@axolot/master-mcp

Master Server.
The full picture.

The Master MCP is the highest-level server. It understands your entire site structure — every page, every slot, every module — and exposes tools that let your AI make meaningful, high-level editorial and architectural decisions.

What your AI can do

listPages()

Get all pages with their slugs, titles, and last modified date.

getSlots(pageId)

Read all editable slots on a specific page, including current values.

updateSlot(key, value)

Write new content to any slot. The AI updates your site directly.

createPage(title, slug)

Scaffold a new page with a slug and register it in the CMS.

getSiteSettings()

Read your site name, logo URL, social links, and contact info.

getActiveModules()

List which modules are enabled and their configuration status.

axolot.mcp.json
"axolot-master": {
  "command": "npx",
  "args": [
    "-y",
    "@axolot/master-mcp"
  ],
  "env": {
    "AXOLOT_SITE_ID": "your-site-id",
    "AXOLOT_API_KEY": "sk-axolot-..."
  }
}

Best for

Content editorsSEO workflowsSite auditsBulk updatesPage creation
Server 02 of 03
@axolot/api-mcp
axolot.mcp.json
"axolot-api": {
  "command": "npx",
  "args": [
    "-y",
    "@axolot/api-mcp"
  ],
  "env": {
    "AXOLOT_API_KEY": "sk-axolot-..."
  }
}

Best for

Infrastructure managementMulti-site agenciesModule activationTeam onboardingSite provisioning

API Server.
Full infrastructure
control.

The API MCP connects directly to the Axolot REST API. Where the Master server focuses on content, this one handles the infrastructure layer — sites, users, modules, and configuration. Perfect for agencies managing multiple clients.

What your AI can do

listSites()

Get all sites in your account with their status and configuration.

createSite(name, domain)

Provision a new client site with isolated storage and credentials.

createBlogPost(title, content)

Publish articles to the Blog module with full metadata control.

createProduct(name, price)

Add new products to the Shop module with variants and images.

activateModule(moduleId)

Enable a module on a site without touching a single line of code.

getSubmissions(formId)

Retrieve and filter form submissions across any form on any site.

Server 03 of 03
@axolot/db-mcp

Database Server.
Direct access.
Full context.

The DB MCP gives your AI direct read and write access to your isolated PostgreSQL schema. No abstraction layer, no guessing — the AI sees your tables, your data, and can run precise queries in the exact context of your project.

What your AI can do

getTables()

List every table in your site schema with column definitions and row counts.

query(sql)

Run any SELECT query within your isolated schema. The AI can explore and analyse your data.

getTableInfo(name)

Get full schema details for a specific table — types, constraints, indexes.

execute(sql)

Run INSERT, UPDATE or DELETE within your schema. With great power, etc.

Schema-isolated by design

The DB MCP can only access your site's specific schema. It has no visibility into other tenants, system tables, or Axolot's internal data.

AI · DB Session
-- User: "Which blog posts have no meta description?"
 db.query(
  "SELECT title, slug
   FROM site_acme.blog_posts
   WHERE meta_description IS NULL
   OR meta_description = ''
   ORDER BY created_at DESC"
)
-- Result: 4 rows
"Why Astro Changed Everything" /blog/astro-changed-everything
"Our Agency Story" /blog/our-story
... 2 more
-- AI: "Found 4 posts. Want me to generate meta descriptions for all of them?"

Best for

Data analysisCustom reportingContent auditsDirect fixesSchema exploration

Quick Setup

Up and running
in two minutes.

Complete configuration

mcp.json
{
  "mcpServers": {
    "axolot-master": {
      "command": "npx",
      "args": ["-y", "@axolot/master-mcp"],
      "env": {
        "AXOLOT_SITE_ID": "<your-site-id>",
        "AXOLOT_API_KEY": "<your-api-key>"
      }
    },
    "axolot-api": {
      "command": "npx",
      "args": ["-y", "@axolot/api-mcp"],
      "env": {
        "AXOLOT_API_KEY": "<your-api-key>"
      }
    },
    "axolot-db": {
      "command": "npx",
      "args": ["-y", "@axolot/db-mcp"],
      "env": {
        "AXOLOT_SITE_ID": "<your-site-id>",
        "AXOLOT_API_KEY": "<your-api-key>"
      }
    }
  }
}

Setup steps

01

Get your Site ID and API Key from the Axolot Dashboard → Settings → API.

02

Add the config block to your AI tool's MCP configuration file.

03

Restart your AI editor. The servers will connect automatically.

04

Ask: "List the pages on my site." You're in.

Compatible with

Cursor Add to ~/.cursor/mcp.json
Claude Desktop Add to claude_desktop_config.json
VS Code Via GitHub Copilot MCP extension
Zed Add to ~/.config/zed/settings.json
Antigravity Auto-configured via Axolot MCP server