Skip to main content

Overview

The Hooked MCP Server (@hooked-so/mcp) lets AI agents — Claude, Cursor, ChatGPT, and any MCP-compatible client — create videos, browse avatars, check status, and discover trends directly from a conversation.
What is MCP? The Model Context Protocol is an open standard that lets AI assistants interact with external tools and APIs natively.

Quick Setup

1

Get your API key

Go to hooked.so → Settings → API Keys and generate a key.
2

Add the config

Paste the JSON config into your AI client (see below).
3

Start creating

Ask your AI to create a video. It handles everything.

Claude Desktop

Add this to your claude_desktop_config.json:
On macOS, the config file is at ~/Library/Application Support/Claude/claude_desktop_config.json

Cursor

Add this to .cursor/mcp.json in your project root:

Available Tools

The MCP server exposes 10 tools:

Video Creation

Resources

Video Management

Example Conversations

Once connected, you can interact naturally: Create a video:
“Create a 30-second product video for my fitness app. Use a young female avatar with an energetic tone.”
Browse resources:
“What AI avatars do you have? Show me the female ones.”
Check status:
“What’s the status of my latest video?”
Get inspiration:
“What’s trending on TikTok in the fitness niche? Create a video based on the top trend.”
Batch creation:
“Create 3 variations of a product ad with different hooks and avatars.”

Tool Details

create_script_to_video

create_prompt_to_video

Environment Variables

Troubleshooting

Ensure npx is available in your PATH. Try running npx @hooked-so/mcp manually in a terminal to check for errors.
Verify your HOOKED_API_KEY is correct and active. Generate a new key at hooked.so → Settings → API Keys.
Restart your AI client after adding the config. Some clients require a full restart to detect new MCP servers.

Next Steps

Skill File

Download the Hooked skill for Cursor and Claude

Webhooks

Receive notifications when videos are ready