What MCP is, for people who do not code
MCP — Model Context Protocol — is the standard that lets AI assistants use external tools safely. Instead of an assistant pretending to know your systems, MCP hands it a typed menu of real actions. VividMotion runs a hosted MCP server at vividmotion.io/api/mcp that turns any MCP-capable assistant — Claude, ChatGPT, Cursor, Codex, Gemini — into an Instagram content studio.
The thirty tools the server exposes
The server exposes about thirty tools: create tools for each design system (editorial, bold poster, full-bleed video, voice-over cards and more), a reel planner, voice and photo catalogs, and render-status checks. Authentication is a single API key in a header. There is nothing to install and no local process to babysit — the assistant calls the hosted endpoint, and the heavy rendering runs on VividMotion's infrastructure.
How a request becomes a render
The flow: your assistant reads your brand profile (look, colors, handle, narration voice) so every composition starts on brand; it plans scenes against a catalog of hand-built components; it generates narration with real text-to-speech; and it submits a render job that produces a 1080×1920 MP4. Scene timing is synced to the narration automatically — each beat lasts exactly as long as its line.
One deliberate boundary: publishing stays manual
One deliberate boundary: publishing is manual. The MCP server will compose and render all day, but posting to Instagram happens only from your dashboard, by you. That separation is what makes it safe to hand an AI the keys to your content pipeline — the assistant does the work, you keep the send button.
Setup in about two minutes
Setup takes about two minutes: a free account at vividmotion.io, a key from the connect page, and one pasted snippet in your assistant's connector settings. The full per-client guide lives at vividmotion.io/mcp.