AI Compatibility

SQLite Core & Universal SQL Querying

The .duc file format is built natively as an SQLite database container by design. Rather than wrapping physical-engineering state in opaque binary streams or massive proprietary formats, .duc stores all document data, elements, geometry, symbol blocks, version history, and project charters, inside a standard relational schema.

Because SQL is a universally understood query paradigm for foundation AI models (LLMs, multimodal models, and autonomous agents), AI tools can immediately inspect, query, and mutate project state without having to guess internal structures or rely on fragile custom parsers:

• Direct Schema Inspection: Models can inspect table structures, columns, and relationships using standard relational SQL statements.
• Selective Operations: Read or update specific elements, properties, or project charter requirements without loading or re-serializing the entire document into context memory.
• Zero Structural Guesswork: Rely on standardized relational DDL schemas (elements, checkpoints, version_graph, search_fts) rather than proprietary black-box formats.

For complete details on our database schema and querying capabilities, visit the SQL Documentation and SQL API Reference.

---

Python Adapter Integration (ducpy)

Python is the primary language used globally for data evaluation, data manipulation, machine learning, and AI agent execution.

.duc provides a first-class Python SDK (ducpy), opening all doors for AI foundation models to interact with the file format programmatically:

• Data Evaluation & Analytics: AI models can leverage ducpy alongside popular data analysis libraries (numpy, pandas, shapely) to evaluate spatial geometry and physical constraints.
• Programmatic CAD Generation: AI models can generate or modify CAD project state directly using Pythonic builder patterns (DucStateBuilder, DucElementBuilder).
• Seamless Agent Tooling: LLMs can invoke ducpy methods as native function calls in agentic workflows to inspect, filter, convert, or export .duc documents.

For SDK setup and usage, visit the Python Adapter Documentation and Python API Reference.

---

Unlocking AI-Native Engineering

By combining an SQLite database core with a first-class Python adapter, the .duc format provides a fully open, AI-compatible foundation for next-generation automated physical-engineering tools.