JSON2Video

JSON2Video MCP Server

Official Model Context Protocol server for JSON2Video, the video API for automated video production at scale.

Connect your AI agent (Claude, Cursor, VS Code, Windsurf, Cline…) to JSON2Video and let it produce finished MP4 videos: build a video from reusable templates, add AI voiceovers, subtitles, AI-generated images and video, render it in the cloud and get back a URL you can publish.

Built and maintained by the JSON2Video team. Two ways to use it: connect to the hosted remote server by URL (no install, sign in with your JSON2Video account), or run the local server from the @json2video/cli npm package over stdio.

Listed in the official MCP Registry as com.json2video/mcp.

What you can do

• "Make a 30-second vertical video about our new product with a voiceover and subtitles."
• "Render my Real estate listing template for each of these 10 properties."
• "Create a template for daily news shorts, then render today's three stories."
• "Upload this logo and add it as a watermark to the video."
• "How many credits do I have left?"

Connect by URL (remote server)

The easiest way. Add this URL to your AI app as a custom MCP server:

https://mcp.json2video.com/mcp

The first time, the app sends you to JSON2Video to sign in and approve the connection (" wants to connect to your JSON2Video account", Allow / Cancel). No API key to copy, nothing to install. Renders use the credits of the account you sign in with.

• Claude (claude.ai and Claude Desktop): Settings → Connectors → Add custom connector, paste the URL, then Connect. On Team and Enterprise plans an organization owner adds it for the organization first.

• ChatGPT: Settings → Connectors → Add → Add custom MCP server, paste the URL and choose OAuth. Availability depends on your ChatGPT plan and workspace.

• Claude Code:

  bash

  claude mcp add --transport http json2video https://mcp.json2video.com/mcp

  then run /mcp and authenticate.

• Cursor (~/.cursor/mcp.json):

  json

  {  "mcpServers": {    "json2video": { "url": "https://mcp.json2video.com/mcp" }  }}

• VS Code (.vscode/mcp.json):

  json

  {  "servers": {    "json2video": { "type": "http", "url": "https://mcp.json2video.com/mcp" }  }}

Connected apps. Each app you allow gets its own API key, named <App> (connected app), listed in Dashboard → API Keys. To disconnect an app, open Dashboard → Account → Connected apps and click Revoke: its access and its key are removed. An account can have up to 20 API keys.

No browser sign-in? For servers, CI and scripts, send an API key instead, as Authorization: Bearer YOUR_API_KEY or x-api-key: YOUR_API_KEY:

claude mcp add --transport http json2video https://mcp.json2video.com/mcp \  --header "Authorization: Bearer YOUR_API_KEY"

Limits. A render returns a project ID quickly and the assistant checks its progress (check_status, list_recent_renders), so long videos keep rendering in the background. media_upload on the remote server accepts public URLs only (up to 50 MB); to use files from your computer, upload them in the dashboard or use the local server.

Step-by-step guides and troubleshooting: MCP server documentation.

Run it locally

Use the local server for offline work, CI, editors that only support local (stdio) servers, or to upload files from your disk.

Requirements

• Node.js 20 or later
• A JSON2Video API key. Sign up for free and copy the key from the dashboard.

Installation

Claude Code (plugin)

Install the plugin from this repository. Claude Code asks for your API key and stores it securely:

/plugin install json2video --marketplace JSON2Video/json2video-mcp

On older Claude Code versions, run /plugin marketplace add JSON2Video/json2video-mcp first, then /plugin install json2video@json2video.

Claude Code (manual)

The quickest way, from your project folder:

npm install -g @json2video/clijson2video auth --api-key YOUR_API_KEYjson2video mcp install claude

This adds the server to the project's .mcp.json and installs the create-video skill. Use cursor or vscode instead of claude for those editors. Or add it by hand:

claude mcp add json2video --env JSON2VIDEO_API_KEY=YOUR_API_KEY -- npx -y @json2video/cli mcp run

Claude Desktop, Cursor, Windsurf and other clients

Add this to your client's MCP configuration file (claude_desktop_config.json, ~/.cursor/mcp.json, ~/.codeium/windsurf/mcp_config.json…):

{  "mcpServers": {    "json2video": {      "command": "npx",      "args": ["-y", "@json2video/cli", "mcp", "run"],      "env": {        "JSON2VIDEO_API_KEY": "YOUR_API_KEY"      }    }  }}

VS Code

Add this to .vscode/mcp.json:

{  "servers": {    "json2video": {      "type": "stdio",      "command": "npx",      "args": ["-y", "@json2video/cli", "mcp", "run"],      "env": {        "JSON2VIDEO_API_KEY": "YOUR_API_KEY"      }    }  }}

Restart your client after saving. Full instructions for every editor: MCP server documentation.

Tools

What runs on your machine (local server)

The plugin and the local configs above run npx -y @json2video/cli mcp run. npx downloads the published @json2video/cli package from npm and starts the MCP server locally over stdio. The server sends your requests (movie JSON, template variables, media files you choose to upload) to the JSON2Video API at api.json2video.com, authenticated with your API key. To answer help and example requests it also downloads public reference files (documentation, examples, voice and music lists) from cdn.json2video.com and JSON2Video's public S3 buckets. It sends no data to any other service.

With the remote server, your AI app sends the same requests to mcp.json2video.com, which calls the JSON2Video API on your behalf with the key created for that connection. Nothing is sent to any other service.

Configuration (local server)

Renders use the credits of your JSON2Video account. Free accounts include credits to try it out; see pricing.

Resources

• JSON2Video
• Documentation
• Remote MCP server URL: https://mcp.json2video.com/mcp
• MCP server guide
• Video editing with Claude Code
• @json2video/cli on npm
• Support: the chat on json2video.com

License

MIT