Lara Dashboard MCP Server: Manage Your Laravel CMS

What is the Lara Dashboard MCP server?

Lara Dashboard is an open-source Laravel admin panel, headless CMS and modular starter kit with users, roles, permissions, content, email and a REST API. It ships a built-in MCP server so an AI coding agent can operate an installed site, not only write code about it.

The server runs on your own Lara Dashboard install at a /mcp endpoint once turned on in Settings. There is no separate package to run; enabling it and issuing a token is all the server needs.

What it changes in your workflow

• An agent can draft and publish posts or pages, and generate SEO meta, instead of you editing each field by hand.
• CRM leads, deals and tickets can be triaged from a chat prompt when the CRM module is active.
• Form submissions get reviewed and marked viewed without opening the admin list first.
• Routine jobs such as a daily briefing, a cache clear or checking recent log output can be asked for in plain language.

Where it runs

Who it is for

Best for

• Laravel teams running Lara Dashboard who want an agent to publish content, triage forms and run upkeep tasks on the live site.
• Site owners who want scoped, token-based agent access instead of handing over full admin credentials.
• Developers already using the CRUD generator and modules who want installed modules to expose their own agent tools.

Not for

• Reading a live canvas before any code gets written. When the task is pulling context straight out of Figma files, Figma MCP reaches that canvas instead of a Laravel site.
• Pulling ready-made UI components into a separate front end. When the job is browsing component registries instead of a CMS, the shadcn MCP server is the closer fit.
• Managing a separate self-hosted tool's own shapes and frames. When the target is a self-hosted Penpot instance rather than a Laravel CMS, the Penpot MCP server handles that reading and editing work.

Connect Lara Dashboard MCP Server: Manage Your Laravel CMS to your AI client

git clone https://github.com/laradashboard/laradashboard.git

1. Install and run Lara Dashboard

   Clone the repo, run composer install and npm install, copy .env.example to .env, run php artisan key:generate and php artisan migrate:fresh --seed, then start it with composer run dev.

2. Enable MCP and generate a token

   Open Settings, MCP in the admin panel, turn the server on and save, then click Generate MCP token and copy it. It shows once and differs from a regular API login token.

3. Add the config to your client

   Use the snippet above in your agent client's MCP settings, swapping in your own site URL and the copied token. Settings, MCP includes ready-made JSON for several clients so you can copy instead of typing it.

4. Ask it something

   Try listing recent posts, drafting a new one, or running the daily briefing. The endpoint answers with a 404 until MCP is turned on in settings.

Client configuration

{
  "mcpServers": {
    "laradashboard": {
      "url": "https://your-site.test/mcp",
      "headers": {
        "Authorization": "Bearer YOUR_MCP_AGENT_TOKEN"
      }
    }
  }
}

Endpoints and ports

What you can do with it

• Built into the admin panelThe MCP server ships inside Lara Dashboard itself, so turning it on in Settings is the only setup step beyond issuing a token.
• Scoped agent tokensSettings, MCP generates a dedicated agent token with chosen abilities, kept separate from regular API login tokens.
• Module tool registryInstalled modules register their own tools and briefing providers through the McpFilterHook registry, so CRM or Forms modules extend what an agent can do.
• Content and SEO actionsAn agent can list, create and update posts and pages, and generate SEO meta fields for them.
• Daily operationsDaily briefings, cache clearing and log listing let an agent handle routine upkeep without a human opening the admin panel.
• Media handlingA dedicated upload tool sends media into the library, with signed checks and a size cap on base64 uploads.
• Ready-made client configSettings, MCP includes JSON an agent client can use directly, rather than requiring hand-written configuration.

Tools it exposes

Frequently asked questions