How Can Developers & AI Agents Generate Videos Programmatically via MCP?
Developers and AI agents can trigger asynchronous text-to-video and image-to-video jobs directly from Claude Desktop, Cursor, or custom workflows using OpenVideoGenerator’s native Model Context Protocol (MCP) server at /api/mcp. Authenticate with a Bearer API key from your dashboard, submit jobs across Kling, MiniMax, Luma, or Hunyuan, and poll task status until the clean MP4 URL is ready.
- Zero custom scraping needed—standard JSON-RPC over HTTP /api/mcp endpoint.
- Supports Cursor (.cursor/mcp.json), Claude Desktop, and Antigravity out of the box.
- Uses the exact same unified credit wallet and automatic failure-refund protection.
Try This Workflow Live
Pre-configured with Developer MCP (/api/mcp) and 16:9 framing. Edit the prompt or upload a reference image below.
Options & Trade-Offs Compared
| Option | Best For | Output | Watermark | Cost | Verdict |
|---|---|---|---|---|---|
| OpenVideoGenerator MCP (/api/mcp) | AI Agents, Cursor, Claude Desktop automation | 1080p MP4 | No watermark with credits | Unified credit wallet | Native agent-ready video tool |
| Raw Per-Vendor SDKs | Single-vendor enterprise apps | Varies | Varies | Multiple keys & billing accounts | High integration & billing overhead |
Step-by-Step Execution
Generate an API Key
Sign in and create a Bearer API key inside Settings → API Keys.
Add /api/mcp to Your Agent Config
Point Cursor or Claude Desktop to https://openvideogenerator.com/api/mcp with your Authorization Bearer header.
Dispatch & Poll Async Video Jobs
Let your coding or content agent submit prompts, wait for completion, and retrieve the clean MP4 URL.
Detailed Q&A
Where can I find the full MCP setup JSON?
Both our /llms.txt directory and Settings → API Keys page include copy-pasteable Cursor and Claude Desktop configs.
Does MCP support multiple video models?
Yes, you can specify Kling, MiniMax, Luma Ray-2, or Hunyuan per tool call.