JungleGym Releases

May 23, 2026

May 15, 2026

Adds zero dependency support, introduces LiteRT-LM, and upgrades to Transformers v5.

May 12, 2026

Pascal Editor v0.8.0 brings community-driven improvements and marks the first public npm release of the MCP server package.

May 10, 2026

Patch release fixing a sandbox proxy iframe timeout issue caused by improper cleanup in strict mode.

May 9, 2026

v0.7.0 is a major feature release expanding door/window systems, introducing column nodes, material paint mode, and MCP server support.

May 9, 2026

v2.8.0 focuses on observability and system stability, adding comprehensive Prometheus metrics, Grafana dashboards, and improved OOM handling.

May 9, 2026

This AI tool release adds support for Qwen3.5 and Gemma 4 models, optimizes VRAM usage, and introduces an integrated benchmarking system alongside reproducible run capabilities.

May 5, 2026

A minor UI client update that adds new props to the AppRenderer component to provide host information and capabilities.

May 1, 2026

This release is primarily a bug fix, security, and compatibility update addressing issues with LiteLLM, llama-cpp-python, and Model2Vec, while adding support for last token pooling and streaming Labels pipeline.

Apr 29, 2026

This release focuses on inference engine updates, introducing support for new models like DeepSeek V3.2, Qwen3.6, MiniMax-M2.7, and GLM-5.1. It also includes improvements to replica management, UI optimization, and critical security and Docker build fixes.

Apr 25, 2026

This cluster management tool update adds new CLI commands for remote troubleshooting, including machine logs and round-trip time monitoring. It also introduces nightly builds and improved service log filtering.

Apr 24, 2026

A major feature release for the 3D editor featuring a multi-surface material system, automatic wall-room generation, and robust undo/redo capabilities. It also introduces full curved fence support and WebGPU renderer improvements.

Apr 22, 2026

This screen recording application update focuses on camera reliability and performance enhancements. It introduces an Ultra quality mode, segmented instant mode recordings, and various UI refinements for better user experience.

Apr 15, 2026

This version adds support for Qwen3.5 and Qwen3-TTS models, along with enhanced resilience for model replicas after supervisor restarts. It also includes model JSON updates and fixes for Docker builds and audio processing.

Apr 13, 2026

A minor update that changes the `sessionid` type from integer to string to facilitate future extensibility, requiring corresponding changes to the frontend API interface.

Apr 11, 2026

This release focuses on improving the developer experience of the deployment workflow and CLI. Key additions include pre-deploy hooks for database migrations, enhanced Compose support (context pinning, PID, and shm_size), and making system SSH the default connection method.

Apr 10, 2026

This release represents a major architectural refactor introducing a plugin-based system for digital human models and TTS. It reorganizes code into modular directories, standardizes base classes for inference and audio processing, and adds new output capabilities.

Mar 31, 2026

A major version bump intended to correct a previous semantic release error that failed to properly increment the major version.

Mar 12, 2026

A minor patch release that increments the version to accommodate the removal of legacy code.

Mar 12, 2026