Bring your own AI: Claude, ChatGPT, or anything that speaks MCP. Let it drive your Vū display and content in plain language, with one secure sign-in scoped to your account.
Run your displays, spaces, and content entirely through natural language.
No dashboards, no menus. Describe what you want and your agent does the rest, across every space.
Your telemetry doesn't just report. Agents read it and act, orchestrated through the Vū AI Gateway, so issues resolve before anyone opens a dashboard.
The Vū AI Gateway is the MCP hub at the center. Bring Claude, ChatGPT, or any agent on one side, and reach your displays, media, spaces, and telemetry on the other.
Authentication runs through VŪ Studio itself. No token juggling, no separate accounts. The agent earns exactly your access, nothing more.
“Set up the VŪ Gateway so you can run my displays. Add the MCP server at https://mcp.vu.ai/mcp, then walk me through signing in with VŪ Studio.”
Point your MCP client at mcp.vu.ai and log in with VŪ Studio once. Your agent is now you.
The agent confirms which display it's driving before sending anything, then reuses it for the session.
“Put my latest render on Studio A.” It finds the media, sends it, and verifies it's live.
Use whichever client you already work in.
Claude
ChatGPT
Gemini
GitHub Copilot
Cursor
Windsurf
Perplexity
Zed
claude mcp add --transport http vu-gateway https://mcp.vu.ai/mcphttps://mcp.vu.ai/mcphttps://mcp.vu.ai/mcpReal workflows, not decoration. Your agent drives a display, finds your media, and lights up the whole room.
Ask in plain language and it sends the command, swaps the content, or tunes the look, then reads the display back to confirm it landed.
Describe what you want and your agent searches your uploads, folders, and stock packs, then plays it straight to the display.
Name one display or all of them. Your agent fans the command out across every display and confirms each one.