
VidMCP is a production MCP server for text-prompt mattes, behind-the-subject composition, and education plates — built for Grok, Cursor, and any MCP host.
Flow fields, glowing tesseracts, depth mattes, kinetic type — click any card to play.





Also: agent pipeline · all files in /assets/demos
Tools over stdio — not a thin FFmpeg wrapper with marketing glitter.
Segment and track a subject from language. Mock offline; MLX / SAM when you need fidelity.
Grades, plates, particles composite under the person — never painted on their face.
Math plates and narration hooks for lessons that hold structure.
Plan → perceive → compose → critic. Quality gates instead of one-shot hope.
A sequence an agent can run end-to-end.
Point Grok or Cursor at the VidMCP binary.
Name the subject. Get a tracked alpha matte.
Build the world behind them — not on top.
Composite, review, iterate with the harness.
Start on mock (no GPU). Flip to MLX on Apple Silicon for real mattes.
# clone + run git clone https://github.com/dhirajlochib/VidMcp.git cd VidMcp && python3 -m venv .venv source .venv/bin/activate && pip install -e . export VIDMCP_SAM_BACKEND=mock vidmcp
{
"mcpServers": {
"vidmcp": {
"command": "/path/to/VidMcp/.venv/bin/vidmcp",
"env": {
"VIDMCP_SAM_BACKEND": "mock"
}
}
}
}
Building where agents stop describing media and start producing it. Also: fintech, smart contracts, and AI products — dhirajlochib.com
Mock backend for first run. Real SAM when you’re ready.