Schema TransformationAlpha

Documentation Portal

Documentation

Learn how Schema Transformation Toolkit works, from your first transformation to building custom parsers and generators.

Quick StartPlaygroundGitHub

Transformation pipeline

TypeScript
↓
Shape IR
↓
Rust

Start Here

Pick the path that matches your goal

Choose a path based on what you need right now instead of scanning every section in order.

New to the toolkit

I want the fastest onboarding path

Get the smallest mental model first, then validate a real transformation in the browser.

Quick start

↓

First transformation

↓

Shape IR

Planning an integration

I need to evaluate package-level integration

Understand which package surface to adopt before wiring the toolkit into another system.

Package reference

↓

Loss analysis

↓

Playground validation

Studying internals

I want to understand how the pipeline is built

Follow the pipeline from Shape IR through traversal and architecture boundaries before extending generators.

System architecture

↓

Traversal

↓

Custom generators

Browse By Role

Start from the role you are playing

Use audience metadata to jump straight to the docs that match how you are approaching the toolkit today.

01

New users2 docs
Quick start

Start with the playground for exploration and published toolkit packages for integration.

Suggested reads

Quick start

Run your first transformation

02

Learners2 docs
What is Shape IR?

Shape IR is the shared semantic layer that keeps multi-format transformation work coherent.

Suggested reads

What is Shape IR?

Traversal explained

03

Integrators2 docs
Understand loss analysis

Use diagnostics to understand where semantics degrade or remain unsupported.

Suggested reads

Understand loss analysis

API and package reference

04

Extenders1 docs
Create a custom generator

Map a stable slice of Shape IR into a target format incrementally.

Suggested reads

Create a custom generator

05

Contributors1 docs
System architecture

Understand how parser, IR, traversal, generator, and diagnostics layers stay separated inside one pipeline.

Suggested reads

System architecture

06

Evaluators1 docs
Common questions

Answer the recurring adoption questions around IR, trust, browser support, and performance.

Suggested reads

Common questions

Quick Start

Get productive in under fifteen minutes

Help a first-time visitor become productive in a few minutes instead of making them decode the repo structure first.

2 min

Install

Understand where the web app and published toolkit packages fit together.

5 min

First Transformation

Use the hosted workspace to run a supported transformation without setting up the full toolchain.

10 min

Understand Shape IR

Get the mental model behind Shape IR before diving deeper into parsers and generators.

15 min

Create a Generator

See the direction for custom generators and the architecture that will support them.

Learning Path

Follow a path that matches your depth

The documentation should make it obvious whether you are just starting, integrating the toolkit, or studying the internals.

Beginner

Introduction

↓

Installation

↓

Playground

↓

First transformation

↓

Concepts

Intermediate

Shape IR

↓

Traversal

↓

Parser

↓

Generator

↓

Transformation workflow

Advanced

Architecture

↓

Diagnostics

↓

Performance

↓

Plugin development

↓

Future IR

Documentation Categories

Learn by purpose, not by folder name

Browse the toolkit by intent rather than source folders. Each category should help you decide what to learn next.

Concepts

Understand the mental model behind Shape IR, traversal, parsers, generators, and diagnostics.

• Shape IR

• Traversal

• Parser model

• Generator model

Guides

Follow task-oriented flows that help you install the toolkit, run transformations, and extend the pipeline.

• Install the toolkit

• Run your first transformation

• Extend the workflow

Examples

Jump straight into runnable examples and compare inputs, outputs, and diagnostic behavior.

• Primitive types

• Object schemas

• Enums and unions

• Arrays

Reference

Find the stable APIs, surface areas, and types that matter when integrating the toolkit into real systems.

• Parser API

• Generator API

• Traversal API

• Utilities

Architecture

Understand the internal pipeline, module boundaries, and design choices that keep transformations deterministic.

• Pipeline

• Modules

• Design decisions

• Future IR layers

FAQ

Answer the recurring questions around IR, browser support, loss analysis, and performance.

• Why IR?

• Why not AST?

• Browser support

• Performance

Featured Topics

Start with the highest-signal topics

Start with the ideas and workflows that are most useful right now, not simply the most recently edited pages.

What is Shape IR?

The shortest explanation of why a shared IR matters for cross-format transformations.

Learn Shape IR
Traversal explained

See how traversal enables reuse across parsers, generators, and future analysis layers.

Study traversal
First transformation

Use the live app to understand current supported flows and where loss can still happen.

Try round trip
Loss analysis

Follow how diagnostics and semantic caveats should be interpreted downstream.

Inspect diagnostics

Read Then Try

Take a doc path straight into a live validation flow

Move from reading to validation without having to decide on the next step yourself. Each path pairs one doc page with a concrete playground follow-up.

Path 1

Read doc

Understand the first successful round trip

Start with the shortest guide, then confirm that a basic schema round trip behaves the way the docs describe.

Validation focus

Check one object schema, one array, and one enum.

Read first transformationOpen Playground
Path 2

Read doc

Evaluate where semantics can degrade

Read how lossy cases are surfaced, then compare the generated output and diagnostics in the interactive flow.

Validation focus

Look for warnings, unsupported nodes, and fallback output.

Read loss analysisOpen Playground
Path 3

Read doc

Connect Shape IR theory to the working product

Use the concept page to build the mental model first, then inspect whether the current playground flow matches that model.

Validation focus

Trace source format -> Shape IR -> target format mentally.

Read Shape IROpen Playground

Recently Updated

See what moved most recently

Use structured metadata to find the pages that changed most recently without relying on folder names or guesswork.

IntegratorsIntermediateEditor pick
API and package reference

Use the reference layer to choose the right package surface before diving into implementation details.

Last updated

July 28, 2026

EvaluatorsBeginner
Common questions

Answer the recurring adoption questions around IR, trust, browser support, and performance.

Last updated: July 28, 2026

ContributorsAdvanced
System architecture

Understand how parser, IR, traversal, generator, and diagnostics layers stay separated inside one pipeline.

Last updated: July 28, 2026

Roadmap

Track what is solid and what is next

Use the capability snapshot to see where the toolkit is strongest today and which areas are still being expanded.

Current maturity

Parser

92%

Generator

68%

Traversal

90%

Shape IR

94%

Constraint IR

18%

Value IR

8%

Next milestones

Constraint IR

Deepen the next IR layer without losing the clarity of the current Shape IR core.

OpenAPI parser

Expand supported source formats in a way that still keeps the pipeline understandable.

Generator expansion

Turn more transformation paths into believable public product surfaces.

Community

Stay connected to the project direction

Stay close to the project as the toolkit surface expands. Community links should help users ask questions and track progress.

GitHub

Read source, open issues, and inspect releases directly from the repository.

Open GitHub
Playground

Use the interactive workspace to validate a transformation path before integrating it into your own workflow.

Open Playground
Roadmap

Follow the capability direction and implementation priorities that are shaping the next milestone.

View roadmap