Pallet-o-Crates

Experimental pallet layout generator for mixed carton sets.

Pallet-o-Crates is a Tauri desktop tool for generating pallet layouts, switching into manual 3D editing, saving layout samples, and reusing a sample database to guide follow-up runs. It is meant for planning, iteration, and experimentation, not for claiming a guaranteed global optimum.

00 shows a clean baseline generation pass: five carton groups, identical dimensions, readable result.

Showcase

01: the same baseline carton set opened in manual mode and staged off-pallet to show the default editing workspace.	02: a harder real-world manual case, showing how a difficult arrangement can be refined directly in the editor.
03: follow-up generation that adds only missing cartons without sample guidance. Valid, but visibly chaotic and a fair example of the current heuristic limits.	04: the same follow-up scenario with sample database guidance enabled. Still imperfect, but noticeably calmer than the unguided pass.

What It Does

• generates pallet layouts under footprint, height, weight, support, and stacking constraints
• supports full generation and missing-carton follow-up generation
• lets you switch to manual 3D editing for correction, staging, and experimentation
• saves layout samples as reusable JSON references
• uses sample guidance and template-lock workflow to preserve useful structure between runs
• ships with a multilingual UI

Experimental Status

This project is usable, but it is still experimental.

• It often produces practical layouts for common cases.
• It does not guarantee full optimization or a global best result.
• Difficult mixed cases can still become noisy and may require manual cleanup.
• Sample guidance can improve coherence, but it is an assistive heuristic, not a magic solver.

Treat it as a planning and reference tool, not as an authoritative warehouse optimizer.

Why The Sample Database Exists

The sample database is there to make follow-up generation less blind.

• Without guidance, missing-carton generation can preserve validity while drifting into visibly rough placements.
• With guidance and template lock, the app can often keep more of the original structure and produce calmer follow-up layouts.
• The screenshot pair 03 vs 04 is the intended honest demo: improvement, not perfection.

Languages

The UI currently ships with 138 languages.

• pl and en are currently marked as human-verified in the language picker.
• Most other locales are still machine-draft or AI-assisted and need native-speaker review.
• Translation status badges are visible directly in the language list.

If you are a native speaker and want to help, translation review is one of the highest-value contribution paths in this repo. Start with CONTRIBUTING.md, docs/localization-workflow.md, and docs/i18n-language-status.md.

Platform / Release Status

• If a tagged Windows build is published, the intended end-user entry point is the GitHub Releases page.
• The current automated GitHub release pipeline builds Windows artifacts.
• Other platforms can still be built from source through Tauri, but are not yet shipped as first-class release targets in GitHub Actions.
• Public releases should currently be read as 0.x experimental releases.

Command Basics

There are two different development entry points:

• npm run tauri dev: main desktop app workflow, recommended for real feature testing
• npm run dev: Vite frontend only, useful for fast UI work but not a full Tauri runtime check

The same distinction applies to builds:

• npm run tauri build: desktop package build
• npm run build: frontend production build plus TypeScript and i18n checks

A fuller command map lives in docs/command-reference.md.

Local Run

Requirements:

• Node.js 20
• Rust stable toolchain
• Tauri 2 toolchain

Run locally:

npm install
npm run tauri dev

Desktop build:

npm run tauri build

Frontend-only build and checks:

npm run dev
npm run build
npm run preview
npm run test:regression
npm run check:i18n

Known Limitations

• Layout quality is heuristic, not guaranteed optimal.
• Some difficult mixed-carton or late-stage top-off cases still need manual intervention.
• Sample guidance helps only when the sample set is relevant and clean.
• Translation coverage is broad, but review quality is uneven across languages.

Contributing

Small, focused contributions are welcome.

• bug fixes
• packing-quality improvements
• regression tests and benchmarks
• documentation cleanup
• native-speaker translation review

Start here: CONTRIBUTING.md

Support

Optional: if you'd like to leave a small tip or say thanks, you can do that here:

License

LICENSE