v0.3.0

[0.3.0] - 2024-04-22

Added

• USD asset search integration with search_3d_usd_by_text tool
• Ability to search and load pre-existing 3D models from USD libraries
• Support for custom positioning and scaling of USD models
• Direct model transformation capabilities with the improved transform tool
• Enhanced scene management with multi-object placement

Improved

• Scene object manipulation with precise positioning controls
• Asset loading performance and reliability
• Error handling for model search and placement
• Integration with existing physics scene management

Technical Details

• Advanced USD model retrieval system
• Optimized asset loading pipeline
• Position and scale customization for USD models
• Better compatibility with Isaac Sim's native USD handling

[0.2.1] - 2024-04-15

Added

• Beaver3D integration for 3D model generation from text prompts and images
• Asynchronous model loading with asyncio support
• Task caching system to prevent duplicate model generation
• New MCP tools:
  • generate_3d_from_text_or_image for AI-powered 3D asset creation
  • transform for manipulating generated 3D models in the scene
• Texture and material binding for generated 3D models

Improved

• Asynchronous command execution with run_coroutine
• Error handling and reporting for 3D generation tasks
• Performance optimizations for model loading

Technical Details

• Integration with Beaver3D API for 3D generation
• Task monitoring with callback support
• Position and scale customization for generated models

[0.1.0] - 2025-04-02

Added

• Initial implementation of Isaac Sim MCP Extension
• Natural language control interface for Isaac Sim through MCP framework
• Core robot manipulation capabilities:
  • Dynamic placement and positioning of robots (Franka, G1, Go1, Jetbot)
  • Robot movement controls with position updates
  • Multi-robot grid creation (3x3 arrangement support)
• Advanced simulation features:
  • Quadruped robot walking simulation with waypoint navigation
  • Physics-based interactions between robots and environment
  • Custom lighting controls for better scene visualization
• Environment enrichment:
  • Various obstacle types: boxes, spheres, cylinders, cones
  • Wall creation for maze-like environments
  • Dynamic obstacle placement with customizable properties
• Development tools:
  • MCP server integration with Cursor AI
  • Debug interface accessible via local web server
  • Connection status verification with get_scene_info
• Documentation:
  • Installation instructions
  • Example prompts for common simulation scenarios
  • Configuration guidelines

Technical Details

• Extension server running on localhost:8766
• Compatible with NVIDIA Isaac Sim 4.2.0
• Support for Python 3.9+
• MIT License for open development