MCP Server Documentation

Connect your AI coding agent to So-me.studio using the Model Context Protocol (MCP). This guide covers setup instructions for every major AI agent.

Claude Code Claude Desktop Cursor OpenAI Codex Windsurf VS Code (GitHub Copilot)Gemini CLI

Prerequisites

A So-me.studio account with at least one connected social account. Sign up here.
An API key generated from Settings > API Access in your dashboard.
MCP Endpoint: https://api.so-me.studio/mcp
Authentication: Pass your API key via the X-API-Key header.
Transport: Streamable HTTP (POST only — stateless mode).

Available Tools

Once connected, your AI agent can call these 8 tools:

list_posts

List scheduled/published posts with filtering by status and platform

get_post

Get a single post by ID with full details

create_post

Create a new post targeting any connected platform

schedule_post

Schedule an existing post for future publishing

delete_post

Delete a post by its unique identifier

generate_caption

Generate an AI caption optimized for a specific platform and tone

generate_and_schedule

Generate AI content, create a post, and schedule it in one call

get_analytics_summary

Fetch workspace analytics: posts by status, platform, and trends

Setup Guides

1/7

Claude Code

Anthropic's agentic coding tool in your terminal

Config: CLI command or .mcp.json

1Open your terminal and run the following command:
2Claude Code will now have access to all So-me.studio MCP tools.
3Verify by running /mcp in your Claude Code session to see "so-me-studio" listed.

bash

claude mcp add --transport http so-me-studio \
  https://api.so-me.studio/mcp \
  --header "X-API-Key: YOUR_API_KEY"

Alternative: .mcp.json

json

{
  "mcpServers": {
    "so-me-studio": {
      "type": "http",
      "url": "https://api.so-me.studio/mcp",
      "headers": {
        "X-API-Key": "YOUR_API_KEY"
      }
    }
  }
}

Place this file in your project root for project-scoped access, or at ~/.claude/.mcp.json for global access.

2/7

Claude Desktop

Anthropic's desktop application for Claude

Config: claude_desktop_config.json

1Open Claude Desktop and go to Settings > Developer > Edit Config.
2Add the following to your claude_desktop_config.json:
3Save the file and restart Claude Desktop.
4Claude will now have access to So-me.studio tools in your conversations.

json

{
  "mcpServers": {
    "so-me-studio": {
      "url": "https://api.so-me.studio/mcp",
      "headers": {
        "X-API-Key": "YOUR_API_KEY"
      }
    }
  }
}

Config file location — macOS: ~/Library/Application Support/Claude/claude_desktop_config.json | Windows: %APPDATA%\Claude\claude_desktop_config.json

3/7

Cursor

AI-first code editor built on VS Code

Config: .cursor/mcp.json

1Open Cursor and go to Settings > MCP, or create .cursor/mcp.json in your project root.
2Add the following configuration:
3Save and Cursor will automatically connect to the MCP server.
4Verify by checking the MCP panel — "so-me-studio" should show as connected.

json

{
  "mcpServers": {
    "so-me-studio": {
      "url": "https://api.so-me.studio/mcp",
      "headers": {
        "X-API-Key": "YOUR_API_KEY"
      }
    }
  }
}

4/7

OpenAI Codex

OpenAI's coding agent for your terminal and IDE

Config: ~/.codex/config.toml

1Run the following CLI command to add the server:
2Codex will store the configuration in ~/.codex/config.toml.
3Verify by running /mcp in a Codex session to see "so-me-studio" listed.

bash

codex mcp add --transport http so-me-studio \
  https://api.so-me.studio/mcp \
  -H "X-API-Key: YOUR_API_KEY"

Alternative: config.toml

toml

[mcp_servers.so-me-studio]
url = "https://api.so-me.studio/mcp"

[mcp_servers.so-me-studio.env_http_headers]
X-API-Key = "SO_ME_API_KEY"

When using env_http_headers, set the SO_ME_API_KEY environment variable in your shell with your actual API key.

5/7

Windsurf

AI-powered IDE by Codeium with Cascade agent

Config: ~/.codeium/windsurf/mcp_config.json

1Open Windsurf and navigate to Cascade > MCP settings, or edit the config file directly.
2Add the following configuration:
3Save the file and restart Windsurf.
4Cascade will now have access to So-me.studio tools.

json

{
  "mcpServers": {
    "so-me-studio": {
      "serverUrl": "https://api.so-me.studio/mcp",
      "headers": {
        "X-API-Key": "YOUR_API_KEY"
      }
    }
  }
}

6/7

VS Code (GitHub Copilot)

GitHub Copilot agent mode with MCP support

Config: .vscode/mcp.json

1Create a .vscode/mcp.json file in your project root (or use the Copilot settings UI).
2Add the following configuration:
3VS Code will prompt you to start the MCP server when Copilot is in Agent mode.
4Verify by checking the MCP status in the Copilot chat panel — tools from "so-me-studio" should appear.

json

{
  "servers": {
    "so-me-studio": {
      "type": "http",
      "url": "https://api.so-me.studio/mcp",
      "headers": {
        "X-API-Key": "YOUR_API_KEY"
      }
    }
  }
}

7/7

Gemini CLI

Google's AI agent for your terminal

Config: ~/.gemini/settings.json

1Run the following CLI command to add the server:
2Gemini CLI will store the configuration in your settings.json.
3Verify by running /mcp in a Gemini session to see "so-me-studio" listed.

bash

gemini mcp add --transport http so-me-studio \
  https://api.so-me.studio/mcp \
  -H "X-API-Key: YOUR_API_KEY"

Alternative: settings.json

json

{
  "mcpServers": {
    "so-me-studio": {
      "httpUrl": "https://api.so-me.studio/mcp",
      "headers": {
        "X-API-Key": "YOUR_API_KEY"
      }
    }
  }
}

Place in ~/.gemini/settings.json for global access, or .gemini/settings.json in your project root for project-scoped access.

Supported Platforms

Once connected, your agent can post to all platforms you have linked in your So-me.studio dashboard:

Twitter / XInstagramLinkedInFacebookTikTokYouTubeThreads

Troubleshooting

Authentication failed / 401 Unauthorized

Double-check your API key is correct and active. Make sure the header is exactly "X-API-Key" (case-sensitive). Generate a new key from Settings > API Access if needed.

SSE / GET requests return 405

So-me.studio's MCP server runs in stateless mode using Streamable HTTP transport. Only POST requests are supported. Make sure your agent is configured with HTTP transport, not SSE.

Rate limit exceeded

MCP tool calls are rate-limited based on your subscription plan. Upgrade your plan or wait before retrying. Check your usage in Settings > MCP.

Credits exhausted

Each MCP tool call consumes credits. Check your remaining credits in Settings > MCP and upgrade your plan if needed.

Connection timeout

If your agent times out during connection, increase the timeout setting in your MCP configuration. A value of 30000ms (30 seconds) is recommended.

Need help?

If you run into any issues setting up MCP, reach out to our support team or check the API docs for more details.

Contact Support API Documentation AI Agents Overview

MCP Server Documentation

Connect your AI coding agent to So-me.studio using the Model Context Protocol (MCP). This guide covers setup instructions for every major AI agent.

Claude Code Claude Desktop Cursor OpenAI Codex Windsurf VS Code (GitHub Copilot)Gemini CLI

Prerequisites

A So-me.studio account with at least one connected social account. Sign up here.
An API key generated from Settings > API Access in your dashboard.
MCP Endpoint: https://api.so-me.studio/mcp
Authentication: Pass your API key via the X-API-Key header.
Transport: Streamable HTTP (POST only — stateless mode).

Available Tools

Once connected, your AI agent can call these 8 tools:

list_posts

List scheduled/published posts with filtering by status and platform

get_post

Get a single post by ID with full details

create_post

Create a new post targeting any connected platform

schedule_post

Schedule an existing post for future publishing

delete_post

Delete a post by its unique identifier

generate_caption

Generate an AI caption optimized for a specific platform and tone

generate_and_schedule

Generate AI content, create a post, and schedule it in one call

get_analytics_summary

Fetch workspace analytics: posts by status, platform, and trends

Setup Guides

1/7

Claude Code

Anthropic's agentic coding tool in your terminal

Config: CLI command or .mcp.json

1Open your terminal and run the following command:
2Claude Code will now have access to all So-me.studio MCP tools.
3Verify by running /mcp in your Claude Code session to see "so-me-studio" listed.

bash

claude mcp add --transport http so-me-studio \
  https://api.so-me.studio/mcp \
  --header "X-API-Key: YOUR_API_KEY"

Alternative: .mcp.json

json

{
  "mcpServers": {
    "so-me-studio": {
      "type": "http",
      "url": "https://api.so-me.studio/mcp",
      "headers": {
        "X-API-Key": "YOUR_API_KEY"
      }
    }
  }
}

Place this file in your project root for project-scoped access, or at ~/.claude/.mcp.json for global access.

2/7

Claude Desktop

Anthropic's desktop application for Claude

Config: claude_desktop_config.json

1Open Claude Desktop and go to Settings > Developer > Edit Config.
2Add the following to your claude_desktop_config.json:
3Save the file and restart Claude Desktop.
4Claude will now have access to So-me.studio tools in your conversations.

json

{
  "mcpServers": {
    "so-me-studio": {
      "url": "https://api.so-me.studio/mcp",
      "headers": {
        "X-API-Key": "YOUR_API_KEY"
      }
    }
  }
}

Config file location — macOS: ~/Library/Application Support/Claude/claude_desktop_config.json | Windows: %APPDATA%\Claude\claude_desktop_config.json

3/7

Cursor

AI-first code editor built on VS Code

Config: .cursor/mcp.json

1Open Cursor and go to Settings > MCP, or create .cursor/mcp.json in your project root.
2Add the following configuration:
3Save and Cursor will automatically connect to the MCP server.
4Verify by checking the MCP panel — "so-me-studio" should show as connected.

json

{
  "mcpServers": {
    "so-me-studio": {
      "url": "https://api.so-me.studio/mcp",
      "headers": {
        "X-API-Key": "YOUR_API_KEY"
      }
    }
  }
}

4/7

OpenAI Codex

OpenAI's coding agent for your terminal and IDE

Config: ~/.codex/config.toml

1Run the following CLI command to add the server:
2Codex will store the configuration in ~/.codex/config.toml.
3Verify by running /mcp in a Codex session to see "so-me-studio" listed.

bash

codex mcp add --transport http so-me-studio \
  https://api.so-me.studio/mcp \
  -H "X-API-Key: YOUR_API_KEY"

Alternative: config.toml

toml

[mcp_servers.so-me-studio]
url = "https://api.so-me.studio/mcp"

[mcp_servers.so-me-studio.env_http_headers]
X-API-Key = "SO_ME_API_KEY"

When using env_http_headers, set the SO_ME_API_KEY environment variable in your shell with your actual API key.

5/7

Windsurf

AI-powered IDE by Codeium with Cascade agent

Config: ~/.codeium/windsurf/mcp_config.json

1Open Windsurf and navigate to Cascade > MCP settings, or edit the config file directly.
2Add the following configuration:
3Save the file and restart Windsurf.
4Cascade will now have access to So-me.studio tools.

json

{
  "mcpServers": {
    "so-me-studio": {
      "serverUrl": "https://api.so-me.studio/mcp",
      "headers": {
        "X-API-Key": "YOUR_API_KEY"
      }
    }
  }
}

6/7

VS Code (GitHub Copilot)

GitHub Copilot agent mode with MCP support

Config: .vscode/mcp.json

1Create a .vscode/mcp.json file in your project root (or use the Copilot settings UI).
2Add the following configuration:
3VS Code will prompt you to start the MCP server when Copilot is in Agent mode.
4Verify by checking the MCP status in the Copilot chat panel — tools from "so-me-studio" should appear.

json

{
  "servers": {
    "so-me-studio": {
      "type": "http",
      "url": "https://api.so-me.studio/mcp",
      "headers": {
        "X-API-Key": "YOUR_API_KEY"
      }
    }
  }
}

7/7

Gemini CLI

Google's AI agent for your terminal

Config: ~/.gemini/settings.json

1Run the following CLI command to add the server:
2Gemini CLI will store the configuration in your settings.json.
3Verify by running /mcp in a Gemini session to see "so-me-studio" listed.

bash

gemini mcp add --transport http so-me-studio \
  https://api.so-me.studio/mcp \
  -H "X-API-Key: YOUR_API_KEY"

Alternative: settings.json

json

{
  "mcpServers": {
    "so-me-studio": {
      "httpUrl": "https://api.so-me.studio/mcp",
      "headers": {
        "X-API-Key": "YOUR_API_KEY"
      }
    }
  }
}

Place in ~/.gemini/settings.json for global access, or .gemini/settings.json in your project root for project-scoped access.

Supported Platforms

Once connected, your agent can post to all platforms you have linked in your So-me.studio dashboard:

Twitter / XInstagramLinkedInFacebookTikTokYouTubeThreads

Troubleshooting

Authentication failed / 401 Unauthorized

Double-check your API key is correct and active. Make sure the header is exactly "X-API-Key" (case-sensitive). Generate a new key from Settings > API Access if needed.

SSE / GET requests return 405

So-me.studio's MCP server runs in stateless mode using Streamable HTTP transport. Only POST requests are supported. Make sure your agent is configured with HTTP transport, not SSE.

Rate limit exceeded

MCP tool calls are rate-limited based on your subscription plan. Upgrade your plan or wait before retrying. Check your usage in Settings > MCP.

Credits exhausted

Each MCP tool call consumes credits. Check your remaining credits in Settings > MCP and upgrade your plan if needed.

Connection timeout

If your agent times out during connection, increase the timeout setting in your MCP configuration. A value of 30000ms (30 seconds) is recommended.

Need help?

If you run into any issues setting up MCP, reach out to our support team or check the API docs for more details.

Contact Support API Documentation AI Agents Overview