Skip to Content
APIMCP Integration

MCP Integration

MCP integration is in active development. This page fills out with setup guides for Claude Desktop, ChatGPT, and Meta Muse when the tools ship. In the meantime, the REST API on the rest of these pages covers every action MCP will expose.

What it is

MCP (Model Context Protocol) is the standard for surfacing tools to AI assistants. Scenema’s MCP server exposes every Studio action as a tool an assistant can call: write a script, generate keyframes, render a video, manage entities, export.

Positioning

Scenema is not a model aggregator or a fire-and-forget job runner. Every Studio surface has a dependency graph: script sets voiceover, voiceover sets treatment, treatment sets shots, shots set keyframes. The MCP tools expose that graph so a client LLM can reason about the current state and pick the next action.

Third parties running their own client LLM get the same output quality Scenema’s own frontend does, because the creative rules live inside each MCP tool rather than inside the client’s prompt.

When it ships

We are targeting general availability behind a manual pilot before opening it to every plan. Sign up on the waitlist  if you want early access. When it ships this page will cover:

  • Adding the Scenema MCP server to Claude Desktop.
  • Adding it to ChatGPT Custom Connectors.
  • Adding it to Meta Muse.
  • The full tool surface: names, arguments, expected outputs.
  • Rate limits and billing when MCP tools call generation endpoints.

Until then

The REST API covers every action MCP will eventually surface. See Making Requests for the request shape and worked examples in curl, JavaScript, Python, and Go.

Last updated on
Errors