What problem it solves

Turning videos and image sequences into 3D Gaussian Splats involves frame preparation, camera reconstruction, model training, and publishing the result. OOOSplat aims to make this workflow accessible through a desktop interface while keeping source material and generation work on the local computer.

Key features

• One-click local generation — Select a video or image sequence, project directory, and quality level to run frame preparation, COLMAP reconstruction, Brush training, and PLY publication through the application.

  Source: README

• Local processing and privacy — Reconstruction and training run on the user's computer, with source material, project data, models, and logs stored in the selected local project directory.

  Source: README

• Panorama and transparent input support — The application recognizes stitched, monoscopic, and complete 360°×180° ERP panorama inputs, and it can preserve color and Alpha data from transparent MOV and PNG files.

  Source: README

• Previewing and editing models — Generated models can be inspected in a preview workspace with movement, rotation, cropping, point deletion, inversion, and undo and redo operations.

  Source: README

• Multiple export formats — Results can be exported as model files, portrait or landscape MP4 videos, and single-file offline HTML viewers using the current edited result.

  Source: README

Quick start

Copied verbatim from the project's README.

sudo apt install ./OOOSplat-0.7.1-x64-linux.deb

# 检查所有内置引擎
cargo run --manifest-path src-tauri\Cargo.toml --bin splatstudio -- health

# 读取视频信息
cargo run --manifest-path src-tauri\Cargo.toml --bin splatstudio -- probe "D:\Videos\orbit.mp4"

# 查看抽帧计划但不写出画面
cargo run --manifest-path src-tauri\Cargo.toml --bin splatstudio -- plan "D:\Videos\orbit.mp4" --quality balanced

# 单独抽帧
cargo run --manifest-path src-tauri\Cargo.toml --bin splatstudio -- extract "D:\Videos\orbit.mp4" "D:\Frames" --quality fast

# 运行完整流水线（自动检测并选择 COLMAP GPU 或 CPU）
cargo run --manifest-path src-tauri\Cargo.toml --bin splatstudio -- generate "D:\Videos\orbit.mp4" --projects-root "D:\Splat Projects" --quality balanced

# 图片序列使用同一命令，传入文件夹即可
cargo run --manifest-path src-tauri\Cargo.toml --bin splatstudio -- generate "D:\Photos\object" --projects-root "D:\Splat Projects" --quality balanced

Who it may suit

This project may suit users who want to create Gaussian Splat models from videos or image sequences without assembling and configuring individual reconstruction and training commands. It may also suit creators who need local preview, editing, and export workflows for ordinary or panoramic material.

Topics to discuss

• Which additional camera-reconstruction or training controls would be useful without exposing the complexity of the underlying engines?

• How should the project validate local workflows across different GPU and operating-system configurations?

• Which export, preview, or MCP workflows would provide the most value for community experimentation?

Written by ReadmeX AI from the project's README, source and public GitHub data; it may contain mistakes.

Read the original README https://github.com/ooolabdev/ooosplat#readme

ReadmeX opened this community for a trending project; it is not run by the project. Maintainers can claim it from the community page.