How It Works

Build semantic data models using familiar tools - no technical expertise required. From markdown templates to production-ready schemas in minutes.

1

Design with Obsidian

Use markdown templates in Obsidian

Templater plugin guides you through

Define types, constraints, relationships

No XML or schema knowledge required

2

AI Processing

Upload markdown or CSV files

Multi-agent AI detects types

Infers constraints, clusters fields

Links to SNOMED, LOINC, WikiData

3

Generate & Deploy

XSD schemas for validation

RDF triples for semantic web

SHACL validation shapes

JSON schemas, GraphQL, and more

Start with Simple Markdown

Write data models in plain markdown. Our Obsidian templates guide you through every field, turning your domain knowledge into production-ready schemas.

User-friendly data types (text, integer, date)
Automatic intelligent type mapping
Optional LLM enrichment for semantics
## Column: patient_age
Type: integer
Units: years
Description: Patient age at visit
Constraints:
- required: true
- range: [0, 120]
-> Mapped to XdCount (SDC4)

Then Let AI Build Your Application

From data model to production-ready app in one conversation

Your Data Model is the Perfect AI Prompt

Download your SDC package, paste our conversation starter into any LLM (Claude, ChatGPT, Gemini), and watch it generate a production-quality application - automatically conforming to your exact data model with validation, constraints, and semantic metadata built-in.

Framework Agnostic: Choose Python Reflex, React, Django, FastAPI, or anything else
Standards-Based Validation: Optional sdcvalidator library for full SDC4 structural and semantic validation
Complete Package: XSD schema, XML/JSON instances, HTML docs, SHACL shapes, and ready-to-use AI prompt
Minimal SDC4 Knowledge Required: Domain experts design, AI assistants build, standards ensure quality
README-AI-PROMPT.md (included in every download)
$ unzip customer-data.zip
// dm-*.xsd, dm-*.xml, dm-*.json, dm-*.html...
// README-AI-PROMPT.md
-> Paste README into Claude/ChatGPT/Gemini
-> Attach your schema files
-> Specify your requirements
Production app in one conversation
The AI-Assisted Development Advantage
Traditional Approach:
"Build me a customer database" -> AI invents structure -> 10 iterations -> mediocre result
With SDC:
Formal model + AI prompt -> Standards-compliant structure -> Database mappings -> Validation rules -> ONE iteration -> production quality
The Invisible Excellence:
SDC4 standards work behind the scenes. Your team gets interoperable, validated, semantic-rich apps without knowing SDC exists.

Domain experts design models. AI assistants build applications. Standards ensure quality. Reduced technical debt.

Supports Claude, ChatGPT, Gemini, Copilot, and any LLM-assisted coding tool