Materializer - is a set of conventions to establish a precise line-by-line binding between the specification you write - and the code AI generates.

Iterate on specs free text - having AI mirror changes into the code seamlessly, rather than "ask" it to "make no mistakes" :) in countless prompts.

It explores world where code is a material that is molded by AI on-demand, same way CPU-level bytecode is a material produced by compilers.

Materializer currently uses OpenCode CLI and requires it to be installed with provider and model(s) selected. See the Prerequisites section below.

Hightlights

• CodeLens - provides hyperlinks in VSCode to navigate between sources and docs line-by-line
• Sidebar with Compile / Decompile / Visualize actions

Features

> COMPILE - implement application from your logic in .src.md files

• Create new main.src.md file and write your app spec in free text (see examples in the docs)
• Reference other .src.md files to break down your app into smaller pieces
• Click Compile and have your app implemented in an impl/ subdirectory, with line-by-line bindings from your spec to the code

> DECOMPILE - reconstruct existing code into .src.md pseudocode

• Open an existing codebase (monorepo supported)
• Click Decompile and have it converted into .src.md files with code moved to impl/, with the same line-by-line bindings
• Edit specs generated and click Compile to have changes reflected back into code

> DOCUMENT - generate read-only docs from existing implementation

• Open an existing codebase (monorepo supported)
• Click Document and have it documented into hyperdocs/ directory with the same line-by-line bindings

Prerequisites

# Install OpenCode CLI
curl -fsSL https://opencode.ai/install | bash

# Bootstrap Opencode
opencode
> /login # into provider of choice
> /model # select model(s) to use for code generation