graph2md

Convert Supermodel architecture graphs into markdown documentation.

1 graph JSON -> 1,997 markdown files with frontmatter, FAQs, and Mermaid diagrams

Why

Supermodel generates architecture graphs — rich JSON representations of codebases with nodes (functions, classes, modules) and edges (calls, imports, inherits). graph2md turns these graphs into a directory of standalone markdown files, ready for any static site generator.

Each entity gets:

• YAML frontmatter (title, description, node_type, language, domain, tags, etc.)
• Mermaid dependency diagrams (incoming and outgoing relationships)
• Source code blocks with syntax highlighting
• Auto-generated FAQ sections
• Graph metadata (relationship counts, complexity metrics)

Quick Start

go install github.com/supermodeltools/graph2md@latest

graph2md -input graph.json -output ./content

Flags

Flag	Default	Description
-input	graph.json	Path to Supermodel graph JSON
-output	./content	Output directory for markdown files
-enrichments	./enrichments	Directory for enrichment JSON sidecar files
-config	pssg.yaml	Path to pssg config (updates content path)

Output

For each node in the graph, graph2md generates a markdown file like:

---
title: "AuthService"
slug: "authservice"
node_type: "Class"
language: "TypeScript"
domain: "auth"
subdomain: "core"
tags: ["authentication", "service"]
---

## Description
Handles user authentication and session management.

## Dependencies
```mermaid
graph LR
  AuthService --> TokenStore
  AuthService --> UserRepository
  LoginController --> AuthService

Source

```typescript export class AuthService { ... } ```

## Architecture

Single-file Go program. 2,188 lines. Zero external dependencies.

Reads Supermodel's `APIResponse` JSON format, walks the graph nodes and edges, and writes one `.md` file per entity with full frontmatter and content sections.