Rig Builder is a powerful, standalone environment for managing and executing complex Python script hierarchies across any host application (Maya, Blender, Unreal Engine, etc.).

While initially developed for rigging, it has evolved into a versatile tool for building pipeline utilities, automation scripts, and custom DCC tools through a visual, module-based workflow. By assembling reusable building blocks, you can create anything from complex rigs to production-ready scene management tools.

⚙️ Core Concepts

At its heart, Rig Builder operates on a graph-based hierarchy of modules:

• 📦 Modules: The primary building blocks. Modules can represent anything from a rigging step (e.g., Spine, Limb) to a general utility (e.g., Batch Exporter, Scene Cleanup).
• 🎛️ Attributes: Parameters that define module behavior. Attributes can hold any JSON-compatible data.
• 🔗 Connections: Attributes can be "wired" together using relative path (e.g., /parent/input).
• 🧠 Expressions: They alter attribute values at time of value resolution.
• 🖥️ Host Connectivity: Rig Builder connects to host applications and executes modules inside, bringing the result back in real-time.
• 🚀 Execution: When triggered, modules execute top-to-bottom inside the host application, driving the DCC/Engine via its native API.
• 💼 Workspaces: Isolated environments that encapsulate your script hierarchies, Git-backed history, and dedicated tool settings.

✨ Key Features

• 💼 Workspace Management — Organize your work into isolated projects. Seamlessly switch between different toolsets, rigs, or automated pipelines while maintaining focused module hierarchies and persistent environment settings.
• 🤖 Agentic AI Chat — Enhanced Ollama integration with tool-calling capabilities. The AI can now perform semantic searches across modules, write code, add attributes, and much more!
• 🔍 Semantic Module Indexing — Intelligent search that understands the functionality of your scripts. Uses vector embeddings to find the right modules using natural language queries.
• 🛰️ Automatic Host Detection — Zero-configuration connectivity. Rig Builder automatically detects and connects to running instances of Maya, Blender, and Unreal Engine.
• 📟 Integrated REPL — A powerful, host-aware Python REPL for immediate feedback and interactive debugging within your current workspace.
• 📜 Git-Backed Module History — Built-in version control for every module change. Track every save, view granular diffs, and restore previous versions in seconds.
• 🔄 Native Auto-Sync — Real-time synchronization between the application and your files on disk, ensuring your UI always reflects the latest changes.
• 📝 Responsive Markdown Docs — Author and view module documentation in native Markdown for a modern documentation experience.
• 🔌 Model Context Protocol (MCP) — Built-in MCP server exposing workspace modules, allowing external AI agents (or vscode-like editors) to query, inspect, and modify your modules.

🚀 Quick Start

1. Installation

Clone the repository and run the installation script which will set up a virtual environment and install dependencies:

git clone https://github.com/azagoruyko/rigBuilder.git
cd rigBuilder
install.bat

2. Ollama AI Setup (Optional)

To enable Local AI Assistance and Semantic Search, install Ollama and pull the required models:

1. Download: Install from ollama.com.

2. Setup: Run the installer.

3. Pull Models:

   • For code assistance: ollama pull your-favorite-model (e.g., codellama, llama3).
   • For semantic search (REQUIRED for indexing): ollama pull nomic-embed-text.

4. Login (Optional): If using cloud models, open your terminal and run ollama signin.

   Note on Indexing: Rig Builder automatically indexes your modules using vector embeddings. This allows you to search for modules using natural language (e.g., "how to build a spine").

   The default models can be customized via the settings.json file in your active workspace:

   • ollamaModel: Model for code generation (defaults to gpt-oss:20b-cloud).
   • ollamaEmbeddingModel: Model for semantic search (defaults to nomic-embed-text).

3. Launch

Run Rig Builder using the launch script:

run.bat

4. Host Setup (Connectivity)

Rig Builder features Automatic Host Detection. When you launch a supported host (Maya, Blender, Unreal Engine), it will automatically appear in the Host Manager.

To execute scripts inside a host, you need zmq (or pyzmq) installed in that host's Python environment.

💡 zmq will be automatically installed on the first connection if it's missing (using a non-intrusive local installation).

5. Usage

To get started with building your own modules, take a look at the example.rb module provided in the modules directory. This serves as a primary reference for module structure and usage patterns.

🖥️ Host Requirements

To execute Rig Builder modules inside a host application (like Maya, Blender, or Unreal Engine), the host must meet the following requirements:

• Python: Version 3.6+ available within the host environment.
• zmq (or pyzmq): Required for high-performance communication between Rig Builder and the host.
  • Note: Rig Builder will attempt to automatically install zmq into the host environment on first connection if it is missing.

🧪 Testing

Run tests using pytest from the project root:

pytest test_core.py -v