Architecture

The project is organized around a static-feeling frontend backed by generated data.

GitHub repositories
        |
        v
Worker import/update jobs
        |
        v
SQLite/libSQL database
        |
        v
Worker generate job + extractor.nvim + Neovim
        |
        v
colorschemes + colorscheme_groups tables
        |
        v
publish job triggers frontend deploy
        |
        v
vimcolorschemes.com

Responsibilities

Part	Responsibility
The App	User-facing website, pages, filters, rankings, and preview rendering
The Worker	Scheduled repository discovery, metadata updates, preview data generation, and publish coordination
extractor.nvim	Neovim-based highlight extraction for each colorscheme/background combination
SQLite/libSQL	Shared storage for repositories, job events, colorschemes, and highlight groups

System boundaries

GitHub API calls, plugin installation, Neovim execution, and highlight extraction happen outside the app request path in the Worker.

The database is the contract between the Worker and the app. When a change affects that contract, update the app and Worker together and document the expected data shape.

Where to make changes

Goal	Start Here
Change website UI	The App
Change repository discovery	The Worker
Change preview generation	Code Previews
Change extracted highlight data	extractor.nvim
Understand data movement	Data Flow