Skip to main content
The Hera MCP server lets AI assistants generate motion graphics videos on your behalf. It exposes the same capabilities as the REST API, but through the standardized Model Context Protocol — so your AI tools can call Hera directly.

Prerequisites

Available tools

The MCP server exposes three tools:

Setup

Install with add-mcp

Install the MCP server for all your coding agents:
Open Claude Desktop → Settings → Developer → Edit Config (this opens claude_desktop_config.json), then add the Hera server under mcpServers:
Save the file and restart Claude Desktop to load the Hera MCP.
Replace YOUR_API_KEY with your actual API key.

Example workflow

Once connected, you can ask your AI assistant to generate videos naturally:
“Create a 10-second 1080p MP4 video of a spinning globe with a blue gradient background”
Behind the scenes, the assistant will:
  1. Call create_video with your prompt and output settings
  2. Receive a video_id
  3. Poll get_video until the status is success
  4. Return the download URL

Tool reference

create_video

Creates a new video generation job. Parameters: Output configuration: Returns: video_id and project_url.

get_video

Retrieves the status and download URLs for a video job. Parameters: Returns: video_id, status (in-progress, success, or failed), project_url, and an outputs array with individual status, download URL, config, and error (if any).

upload_file

Uploads a remote file to Hera’s CDN for use as model input. For local files, pass the local path and the tool will return a ready-to-use curl command. Parameters: Supported types: images (PNG, JPEG, GIF, SVG, WebP, BMP), videos (MP4, MPEG, WebM, MOV, AVI, FLV, MPG, WMV, 3GP), audio (MP3, WAV, AAC), fonts (TTF, OTF, EOT, WOFF, WOFF2), and CSV. Max size: 10 MB. Returns: url — the hosted URL to pass to create_video fields like reference_image_url, reference_image_urls, reference_video_url, or assets.
Looking for support or have questions? Email us at [email protected].