Skip to content

4dgspl.at — 4D Gaussian Splatting changelog

io.github.snmzzcenk/4dgspl

3 changes recorded. MCP Harbor re-reads this listing from the official registry every six hours and asks the live server what it exposes, and keeps what changed between one look and the next.

  • AGENTS.md +5 −2

    https://raw.githubusercontent.com/snmzzcenk/4dgspl-tools/HEAD/AGENTS.md

    −    - Put the important motion in the **first 4 seconds**: the processing server currently uses the first 4 s at 15 fps.−    - Without `--gonder` it only uploads, so the user can review and submit in the dashboard.+    - Put the important motion in the **first 10 seconds**: the processing server currently uses up to the first 10 s at 15 fps.+    - Without `--gonder` it only uploads, so the user can review and submit in the dashboard. No calibration file (real footage only): add `--colmap` and the server tries to estimate it from the images (8+ cameras at most 45° apart, textured scene; 6 or fewer usually fail).+ ## Agent skill+ For a GLB 3D animation, video or still image from the same scene as well as 4DGS, follow [`skills/4dgs-blender/SKILL.md`](skills/4dgs-blender/SKILL.md).+ - Spin the model on a turntable or rotate it to show every side; the camera ring already provides the angles. Only the requested motion, body broadly facing cam00.
  • llms.txt +4 −3

    https://4dgspl.at/llms.txt

    − - Input: synchronized videos from 4–8 (or more) cameras, named cam00.mp4, cam01.mp4 …, plus camera calibration. Calibration is uploaded as a zip with COLMAP cameras.txt + images.txt (PINHOLE or SIMPLE_PINHOLE model, image names cam00.jpg … matching the video names) or as 4dgs-cameras JSON (OpenCV convention). Automatic calibration from the images with COLMAP on the server is in development.− - Processing length: the server currently processes the first 4 seconds of each capture at 15 fps (60 frames).− 1. Build or import the model in Blender (metres, origin, pure black world, matte materials; key motion in the first 4 seconds).+ - Input: synchronized videos from 4–8 (or more) cameras, named cam00.mp4, cam01.mp4 …, plus camera calibration. Calibration is uploaded as a zip with COLMAP cameras.txt + images.txt (PINHOLE or SIMPLE_PINHOLE model, image names cam00.jpg … matching the video names) or as 4dgs-cameras JSON (OpenCV convention). Calibration is required: without a file the project cannot be submitted, unless the user explicitly ticks "Create it from the images (COLMAP)" on the project page or an agent passes --colma+ - Processing length: the server currently processes up to the first 10 seconds of each capture at 15 fps (up to 150 frames).+ 1. Build or import the model in Blender (metres, origin, pure black world, matte materials; key motion within the first 10 seconds). Do not spin the model or add a turntable rotation: the camera ring provides the angles.+ - [Prompt to render (English)](https://4dgspl.at/en/prompt-to-render.html): generates a Blender prompt for Claude Code or Codex that produces a GLB 3D animation, video, still image and 4DGS videos from one scene; previews GLB, video and images in the browser without uploading, with VR for GLB

  • server.json 2 fields changed
    package_identifier
    https://github.com/snmzzcenk/4dgspl-tools/releases/download/v0.1.0/4dgspl.mcpb https://github.com/snmzzcenk/4dgspl-tools/releases/download/v0.1.1/4dgspl.mcpb
    version
    0.1.0 0.1.1