Overview
The Rooli MCP server exposes your social workspace as a set of tools any MCP-compatible AI can call. Instead of switching tabs, you describe what you want in plain language and your assistant creates drafts, schedules posts, uploads media and reads analytics through a secure, token-scoped connection.
Server URL
https://production-api.rooli.co/api/v1/mcpWhat you can do
Create & schedule posts
Draft content and queue it across your connected channels.
Plan campaigns
Group related posts into campaigns and manage them together.
Manage media
Pull images and video into your library from a URL.
Read analytics
Ask for performance summaries and calendar metrics.
Quick start
Create an MCP token
Generate a secure token from the AI Social Manager. It authorizes your AI to act on a single workspace.
Add Rooli as a connector
Paste the server URL into your AI client's connector settings and name it Rooli, then authorize with your token.
Start prompting
Ask your assistant to draft, schedule and publish content — it calls Rooli's tools on your behalf.
Connect your AI
Remote connector (no-code)
In clients that support remote MCP connectors (Claude, ChatGPT and others), open connector settings, add a new connector named Rooli, paste the server URL and authorize it with your MCP token.
https://production-api.rooli.co/api/v1/mcpConfig file (for developers)
For Claude Desktop or any AI desktop client that reads an MCP config file, add the Rooli server and replace YOUR_MCP_TOKEN with a token you create.
{
"mcpServers": {
"rooli": {
"url": "https://production-api.rooli.co/api/v1/mcp",
"headers": {
"Authorization": "Bearer YOUR_MCP_TOKEN"
}
}
}
}Example prompts
“Draft a launch tweet for our new feature and schedule it for 9am tomorrow.”
“List my connected social profiles and show last week's top posts.”
“Create a 5-post campaign teasing our product launch across X and LinkedIn.”
“Upload this image and attach it to a draft about our team offsite.”
“How did my posts perform this month? Summarize the analytics.”
FAQ
Is it secure?
Every request is authenticated with a token scoped to a single workspace. You choose the permissions when creating it, and you can revoke a token at any time from the AI Social Manager.
Which AI clients are supported?
Any MCP-compatible client — including Claude (Desktop & Code), Cursor, ChatGPT custom connectors, and VS Code extensions like Cline.
What can the AI access?
Only the tools listed above, limited to the scopes on your token. It cannot read your password or any data outside the authorized workspace.
How do I stop access?
Revoke the token from the AI Social Manager. The connection stops working immediately.