Free skill for your AI assistant

Camera-Assisted Projection Mapping

Build masks and fit videos to your projection.

Use a camera to see where the projector’s light lands on your house, stage, or installation. Trace the surfaces, build a layered template, and fit your videos to it. Start with an existing projection map or make a new one.

Get the skill on GitHub Read the instructions
Using a camera to check a projected patternThe projector displays a test pattern. The camera shows where its points land on the surface, giving the assistant measurements it can use to fit the artwork.ProjectorCameraSurfaceTest patternCamera view
Project a test pattern. Photograph where it lands.

What you need

Give your assistant a camera view.

Your assistant needs to inspect images from a camera stream, a camera app, or photographs you provide. It also needs tools to edit your files and, for live tests, send images to the projector.

The skill gives the assistant projection mapping instructions. You supply the camera, projector, editing software, and access to them.

What it does

  1. 01

    Look through the camera

    Get a clear view of the area you want to project onto. Zoom in on uncertain edges and check whether a line is part of the structure or just a shadow.

  2. 02

    Measure where the light lands

    Project a test pattern and find its points in the camera image. Map walls, projecting columns, and other surfaces separately when they sit at different depths.

  3. 03

    Fit the artwork

    Find your existing projects and masks, close broken outlines, and resize or warp the artwork to fit. Keep the layers so you can edit them later.

  4. 04

    Try a short video first

    Play it through the device you use for the show. Take another camera image, check the edges, and fix the map before converting the rest of your videos.

The skill covers gaps in outlines, black masks with transparent openings, resolution changes, and videos that already have an old map baked into them. It keeps camera movement separate from projector movement, so you can work out what actually needs to be remapped.

It can follow an existing Digital Press Works, Photoshop, or After Effects workflow. For an xLights show using Falcon Player, it checks how sequences are paired with videos and tests through the player connected to the projector.

GitHub

Install it in your assistant.

Ask your assistant to install the skill from GitHub. For manual installation, clone the repository and copy the camera-assisted-mapping folder into your agent’s skill library. Keep the folder intact; the instructions link to the guides inside it.

In Codex, start with

$camera-assisted-mapping
Use my camera to map the projection. Find my existing artwork and masks, fit them to the new map, and check a test video before converting the rest.

Tell it where your files are, how to access the camera, and what plays the videos. Say whether it can move the camera or project test patterns.

Keep camera credentials and private footage in your own project.

Check the finished map on the actual surface before running a show. You can prepare files offline, but you still need to see how they land through your projector and player.

Get the skill on GitHub Back to work