Castudio MCP
Connect your AI assistant to Castudio
Plan content, prepare drafts and check results from the AI client you already use. MCP, the Model Context Protocol, connects your assistant to your Castudio workspace.
Included in all paid plans. Your existing plan limits apply.
Your content workflow, connected
Plan and draft
Read your calendar, create platform-specific drafts and attach media already in Castudio.
Review and schedule
Submit drafts for review. With publishing permission, an admin’s assistant can approve, schedule and unschedule posts.
Check publishing status
See each platform’s result, including failures, pending delivery and anything that still needs your action.
Read available metrics
Review stored post results and historical publishing-time observations, with coverage and freshness included.
MCP follows Castudio’s support for LinkedIn, Bluesky, YouTube, TikTok, Facebook, Instagram, Threads and X. Platform access restrictions and supported formats still apply. Check current platform availability.
Before you connect
You need a verified Castudio email address and a workspace whose owner has an active paid plan. Paid-plan trials are included. There is no separate MCP fee.
Your AI client must support remote MCP connections. Its own plan, workspace permissions and security settings may affect access.
Start with read, drafts and metrics access. Publishing is off by default. You choose the workspace and permissions when Castudio asks you to allow the connection.
Connect with OAuth
ChatGPT
- In ChatGPT on the web, open Plugins, choose the plus button, then Add custom MCP server.
- Name it Castudio MCP, paste the server URL above and select OAuth.
- Under Advanced OAuth settings, use Dynamic Client Registration (DCR). You do not need to copy a client ID or secret.
- Select
mcp:read,mcp:draftsandmcp:metrics. Leavemcp:publishunchecked for a connection that prepares content for review. - Review the custom-server notice, choose Create as a plugin and connect. Sign in to Castudio, select your paid workspace, review the permissions and choose Allow connection.
- Use Try in chat, or select your installed Castudio MCP plugin with @ in a conversation.
The ChatGPT connection has been tested with Castudio. This creates a personal connection; Castudio is not listed in the public ChatGPT directory. OpenAI’s custom MCP guide.
Connect with OAuth
Claude
- Open Customize → Connectors, then + Add → Add custom connector.
- Enter Castudio MCP and the server URL above, then continue to the authentication settings.
- Choose Sign in now and, under OAuth client, Register automatically. Castudio uses dynamic registration.
- Add the connector, sign in to Castudio and allow the workspace permissions you need. Enable the connector from the conversation’s connectors menu.
On managed workspaces, an owner may need to add the connector first. These steps follow Claude’s custom connector documentation; this client has not yet been independently tested with Castudio.
Connect with OAuth
Cursor
Add this server to ~/.cursor/mcp.json for all projects, or .cursor/mcp.json for one project. If the file already has servers, add the castudio entry under mcpServers without replacing them.
{
"mcpServers": {
"castudio": {
"url": "https://api.castudio.io/api/mcp"
}
}
}Open Cursor’s MCP settings, connect the Castudio server and complete the OAuth sign-in. Select your paid workspace and permissions in Castudio.
These steps follow Cursor’s MCP documentation; this client has not yet been independently tested with Castudio.
Other AI clients
For a client with remote MCP OAuth and dynamic registration, use the same server URL and complete Castudio sign-in. Select Streamable HTTP if it asks for a transport.
Connecting with an API key
- Sign in to Castudio, select your paid workspace and open AI connections.
- Name a new connection and choose its permissions. Save the key when it appears; it is shown only once.
- Add the server URL to your client. In its private connection settings or secret store, set the
Authorizationheader toBearer YOUR_CASTUDIO_KEY, replacing the placeholder with your key.
Keep keys out of chat prompts and shared files. The exact configuration format depends on your client.
Choose what your assistant can do
Your current workspace role and paid-plan eligibility are checked on each request. Existing platform consent requirements still apply.
Use the metrics Castudio can provide
Your assistant can read the latest stored results for posts published through Castudio. Each response includes coverage and freshness, so it can distinguish a reported zero from unavailable data.
Date filters select posts by publication date. The results are lifetime snapshots for those posts, not activity totals or growth within the selected period. Metrics differ by platform, and YouTube reports remain separate.
Historical publishing-time suggestions need enough observations from a supported account. They describe past results and do not promise future performance. Ask your assistant to state missing data and coverage when summarising results.
Manage or remove a connection
Open AI connections in your Castudio workspace to review and revoke access. Connections expire after 90 days. Revoking a connection stops future MCP requests; posts already scheduled remain in the planner until you unschedule them.
If a connection fails, check that your email is verified, the selected workspace has a paid plan and your client allows custom MCP servers. An expired or revoked connection needs a new sign-in or key.