PYTHON PACKAGE · LOCALHOST INTERFACE

Run your own workspace.

Define a task in YAML. Inspect generated models and measured results. Continue from an earlier design.

Get started

The MVP runs on Python 3.11, Linux or WSL2, Git, and Docker. PyPI publication is pending. Clone the repository to install from source; building the interface also requires Node.js 22 and uv.

git clone https://github.com/Cobeml/Da-Vinci.git
cd Da-Vinci

GitHub repository ↗ · Installation guide ↗

Follow the installation guide, then create and start a workspace:

davinci init my-project --template sensor
cd my-project
davinci setup --template sensor
davinci doctor
davinci serve

Set OPENAI_API_KEY in your environment or workspace .env. Local storage works without MongoDB Atlas.

Guides

YAML and workspace settings ↗

Live runs and continuation ↗

Custom Python tasks and coding-agent prompt ↗

Reflection, memory, tools, and independent evaluation ↗

MongoDB Atlas for Scalable Model Improvement ↗

Development and release ↗

CadQuery geometry and physics estimates are evaluated against the selected task contract. Results depend on its assumptions and fidelity.