# OpusClip [Health: Active]

**Category:** 🎥 Multimedia Process  
**Repository:** https://github.com/opus-pro/opusclip-mcp  
**GitHub Stars:** 8  
**npm Downloads (last month):** 287  
**Views:** 1  
**Installs:** 0  
**Upvotes:** 0  
**Directory Page:** https://allmcps.com/mcp/opusclip

## Description
Turn long videos into AI-curated short clips: caption, reframe, thumbnail, schedule, and publish.

## Claude Desktop Quick Installation
Install path detected from listing signals. Uses `npx` (confidence: high):

```json
"mcpServers": {
  "opusclip": {
    "command": "npx",
    "args": ["-y","@opusclip/mcp"]
  }
}
```

## Documentation

## What OpusClip MCP server does

The OpusClip MCP server exposes OpusClip’s video creation and publishing workflow to MCP-compatible clients. An agent can submit a video file or URL and request AI-selected moments that are scored for virality. The resulting clips can be edited with operations such as trimming, reframing, layout changes, and caption styling.

The service also supports transcript retrieval, thumbnail generation, and social copy creation. Social copy can include titles, descriptions, and hashtags for each platform. Clips can be grouped into collections, exported, posted, or scheduled to connected social accounts. Usage tools provide information about remaining credits and the current plan.

## How it works

OpusClip runs as a hosted MCP endpoint at `https://mcp.opus.pro/mcp`. Clients communicate with it over Streamable HTTP, while the provided npm and PyPI launchers proxy stdio connections to the hosted service through `mcp-remote`.

Authentication uses OAuth in the browser. On the first tool call, the client opens a browser window where the user signs in to an OpusClip account and approves access.

## Setup and configuration

For clients that support remote MCP connections, add the endpoint as a server URL:

```json
{
  "mcpServers": {
    "opusclip": {
      "url": "https://mcp.opus.pro/mcp"
    }
  }
}
```

For stdio-only clients, use the npm launcher with `npx -y @opusclip/mcp`, or use the Python-oriented `uvx opusclip-mcp` launcher. The Python launcher also requires Node.js 18 or newer on the system path because both launchers rely on `mcp-remote`.

Claude.ai and Claude Desktop can add the endpoint as a custom connector. Cursor supports a new MCP server in its settings, Cline accepts the configuration in `cline_mcp_settings.json`, and Claude Code can register the server with `claude mcp add --transport http opusclip https://mcp.opus.pro/mcp`.

## Tools and capabilities

The OpusClip MCP server covers these workflows:

- Submit a local video file or URL for AI clip curation.
- Edit clips with trimming, reframing, layouts, and captions.
- Retrieve transcripts and create styled captions.
- Generate thumbnail options.
- Produce titles, descriptions, and hashtags for social platforms.
- Organize clips into collections and export them.
- Publish immediately or schedule posts to connected social accounts.
- Check account credits and plan usage.

## Limitations and notes

This is a hosted service rather than a package that the user builds and runs locally. OAuth access and an OpusClip account are required on first use. Scheduling and publishing depend on connected social accounts, while clip generation and other actions use the account’s available credits. The README points to OpusClip’s API reference for the complete tool definitions and supported options.

_Full upstream README: https://allmcps.com/mcp/opusclip/readme_