AI-Assisted Setup

Chanterelle ships with a built-in AI skill that teaches AI coding assistants — GitHub Copilot, Cursor, Windsurf, and others — how to scaffold and configure Chanterelle projects, including model_meta.json, interactive.json, analytics.json, and handler_io.py.

This means you can describe what you want in natural language, and your AI assistant will generate valid, well-structured project files for you.

How It Works

The Chanterelle repository includes a chanterelle-project skill folder under .github/skills/. It contains:

• SKILL.md — the main entry point that AI assistants read, with the scaffolding procedure and key constraints
• references/model-project.md — full model_meta.json schema and handler_io.py function signatures
• references/interactive-project.md — interactive.json schema and interactive handler_io.py pattern
• references/analytics-project.md — analytics.json schema
• references/visualization-types.md — all supported visualization items (tables, charts, Plotly, Markdown, images, etc.)

When the skill is in your workspace, AI assistants automatically pick it up and use it to:

• Scaffold a new project — create the right folder structure and files for model, analytics, or interactive projects
• Write model_meta.json — define inputs, outputs, presets, and UI groupings with correct types and structure
• Write interactive.json — configure agent metadata and Python environment settings
• Write analytics.json — build sections with tables, charts, Plotly visualizations, Markdown, and more
• Generate handler_io.py — produce the correct function signatures (model_fn, predict_fn, initialize, on_input, etc.) with proper return formats

Supported Editors

Any AI coding assistant that reads project-level instruction files:

• VS Code with GitHub Copilot
• Cursor
• Windsurf
• Other AI-enabled editors and IDE plugins

Quick Start

1. Download the skill folder and unzip it into your workspace so the structure is .github/skills/chanterelle-project/
2. Open your project folder in your editor with your AI assistant enabled
3. Ask the assistant to create a project:

Example prompts:

"Create a Chanterelle model project for an iris classification model with 4 numeric inputs and a category output"

"Scaffold an interactive project with a conversational SQL agent"

"Build an analytics dashboard with a bar chart and a summary table"

The assistant will generate the correct files with valid schemas, ready to open in Chanterelle.

What It Covers

Project Type	Files Generated	What the AI Handles
🧠 Model	model_meta.json, handler_io.py, model_findings.json	Input/output definitions, Python function signatures, visualization sections
📊 Analytics	analytics.json	Sections, items, chart configs, Markdown content, $href references
💬 Interactive	interactive.json, handler_io.py	Agent metadata, initialize() and on_input() functions, dynamic form returns

Tips

• The AI knows all 9 input types (float, int, string, category, boolean, textarea, file, button, yes_no) and will pick the right one based on your description
• It understands the visualization system — ask for specific chart types, tables, or Plotly specs and it will produce correct item formats
• For interactive projects, it knows that module-level Python variables persist across turns (the process stays alive)
• You can iterate — ask the assistant to modify or extend generated files, and it will respect the existing structure