MCP tools
Connect an MCP-compatible assistant to Scenvy and use the same verified production workflow through tools instead of a browser.
MCP is useful when Scenvy is part of a larger research or publishing workflow. The server exposes the capabilities allowed by your key, while research, rights checks and final review still happen inside Scenvy.
Connect a client
Create an MCP key at MCP, copy the server URL shown there, and add it to your client as an HTTP MCP server:
{
"mcpServers": {
"scenvy": {
"url": "https://your-scenvy-api/v1/mcp",
"headers": {
"Authorization": "Bearer YOUR_SCENVY_MCP_KEY"
}
}
}
}The exact URL is always the one in your developer console. Do not put the key in a prompt, browser bundle, public repository or shared config file.
What the tools do
| Tool | Use it for |
|---|---|
create_video | Create and save a production brief; does not start the full run |
create_editor_project | Create a standalone editor project |
create_map | Create a map composition or queue a render |
create_visual | Create a visual module composition or queue a render |
generate_thumbnail | Generate thumbnail variants using reference video IDs |
Start with initialize, call tools/list to see the tools available to your key, then use tools/call. Tool names and input schemas may change as capabilities are added, so the live tool list is the source of truth.
Starting a production, checking jobs, editing timelines and downloading results use the corresponding REST endpoints. These are not additional MCP tools.
Good operating patterns
- Use an Idempotency-Key for create_video or create_editor_project retries; other tools do not support it.
- Keep returned IDs and poll the REST status URL for queued renders instead of creating another job.
- Let Scenvy finish research and rights checks before asking for a script or render.
- Request only the scopes your workflow needs; revoke a key immediately if it leaks.