Art Boards
Tools for coordinating agent sessions and discovering and visualizing resources and artefacts.
Welcome toNoodlestan
A home for software projects, experiments, crazy ideas, and things worth building.
Projects
Tools for working with code and metadata, component libraries and design systems, AI agent tooling, documentation infrastructure, and creative experiments.
The projects are independent, but many are connected and built to work together.
No projects match
Art Domains
Distributable artificial resources within knowledge and work domains: structures, templates, processes, agent instructions, modes, and skills.
Art JS
Libraries and tools for working with Art MD as a programming language, including records, resources, modules, programs, program builders, validators, projectors, language services, and related tooling.
Art Lib
Shared libraries and tools for building CLIs and Tools.
Art MD
Language specification and JavaScript libraries for extracting structured data into a MDAST derived AST, with an open construct registry. Provides parsing, validation, transformation, and serialization of Art MD content.
Art Works
Tools for discovering projects, managing knowledge and context resources, and operating across multiple checkouts: cloning, visualizing, synchronizing, and executing commands.
Artificials
A collection of tools and resources to generate and manage knowledge artefacts and agent instructions.
No Comply
SolidJS layered library ecosystem spanning foundational types, accessibility primitives, contexts and services, composable controllers, a themeable component library, and developer tooling.
Noodlestan Conventions
Conventions packages for Noodlestan stack, including Typescript, JSX, Solid JS, and SCSS.
Noodlestan Web
Repository of Noodlestan websites — typically built with Astro.
Purrception
Collection of abstractions and tools for building metadata extraction pipelines from local and remote sources using heuristics, schemas, and factories.
Purrfect
Collection of composable tools and abstractions for building metadata extraction pipelines, aggregations, indexes, visualizations, and metadata endpoints.
Purrpose
Single-purpose utilities for in-browser transformations: JSX/TSX compilation, SolidJS JSX transform presets, and Shiki-based syntax highlighting.
Purrtrait
Collection of tools for rendering code, visualizing metadata, and creating representations for humans and tools.
Work Flows
Noodlestan projects are built in the open, one small piece at a time, many moving parts evolving together.
The projects live in separate repositories and are coordinated through a shared workspace. Each repository can be built and used on its own.
Work is organised through roadmaps, milestones, plans, and iterations. Plans describe the work, its scope, commits, and how completion will be verified. Roadmaps and milestones coordinate priorities and dependencies.
Projects also carry structured records (Art) describing what they are and how they relate to the rest of the ecosystem. These records are used as sources of truth by agents as well as by tools.
Artificial Work
AI agents are used extensively for planning, drafting, coding, documentation, and automating parts of the work. Since part of the work is to make agentic development more effective I have been able to keep the bill under $20 bucks a month.
The motivation of all this artificial work is to make the work more effective and to turn its results into reusable tools, conventions, and infrastructure. It operates under the principle that for every artificial process, there should be a plan to decompose it and extract every deterministic operation into something that can be made explicit, reproducible, and independent of the agent performing it.
The AGENTS, skills, and other types of context files (a lot MD and a little Art here and there) are not public yet because the goal is to distribute them in different formats so they can be used in different platforms without the Noodlestan dependencies.
The tools, conventions, and ways of working are themselves part of the experiment.
Contributing
About
Hi, I am André Torgal.
Noodlestan is a land of dreams, experimentation, and creativity.
For the past few years I have been keeping my favourite projects and ideas here, without any real plan or particular goals. Writing software because. Just because.
Meanwhile, I decided it was time to publish everything properly.
So I have been working towards stable, documented releases of everything listed here, and making the whole thing easier to discover, use, understand, and contribute to.