Skip to content

AI Playground Version v3.2.0 Beta

Latest
Latest

Choose a tag to compare

View all tags
@qiacheng qiacheng released this 28 Aug 20:45
· 2 commits to main since this release
v3.2.0-beta-rc2
f2379c7
This commit was created on GitHub.com and signed with GitHub’s verified signature.
GPG key ID: B5690EEEBB952194
Verified
Learn about vigilant mode.

AI Playground Version v3.2.0 beta (rc2)

AI Playground 3.2.0 beta is a major update from 3.1.2 beta, with model management, Agent Mode and game presets, expanded Home Agent and speech options, cloud mode, unified history, and broad UI and backend improvements.

Version 3.2.0 beta updates

  • Model management — Models view to browse, search, filter, favorite, download, edit capabilities, reveal in folder, and permanently delete models; model folder paths editor; speech models visible for pre-download
  • Agent Mode & game presets — Agent Mode harness for tool-enabled agent presets; Game Agent (plan, edit, and play-test HTML/JS games, cover art, games library); Quick Coder one-shot game creation with handoff to Game Agent; option to hide games from the library; reset-to-default for new agentic presets
  • Home Agent: local web channel — LAN-accessible web chat (port, optional password), streaming replies, conversation history, file uploads, and image/video/audio/document rendering, in addition to Telegram and Slack
  • Cloud mode — Run chat and agentic flows against external cloud providers instead of local models; improved PDF and text document handling in cloud mode
  • Speech — Kokoro text-to-speech preset (OpenVINO); expanded speech-to-text and text-to-speech presets and settings; Qwen3-TTS as selectable preset and agent tool; TTS/STT backends start only when needed
  • Chat & presets — Chat presets merged and selected from the main mode bar; contextual Help mode (? toggle with preset-specific topics); unified history (one list for chat, agent sessions, audio, and generated media with global search and day grouping); context usage progress bar; current preset and model shown in the UI
  • Pascari aiDAPTIV — Knowledge Manager RAG with shared-prefix KV-cache reuse for improved RAG time-to-first-token on supported Pascari aiDAPTIV SSD offload configurations
  • Hardware & performance — User-selectable default GPU/device with UUID-based identification; OpenVINO OVMS update with INT4 KV cache; ComfyUI on Linux probes XPU before launch; PyTorch updated to 2.13.x line
  • Installer — All-user installer with shared model and backend folder to reduce permission issues on multi-user PCs
  • MCP — Fixes and new MCP info panel for Agentic Chat
  • Reliability & fixes — Startup and chat-history stability; reasoning display while in progress; ReActor face-swap model install path; PDF parsing migrated to unpdf; Hugging Face Xet download progress reporting; cleaner settings layout; security and dependency updates; Linux backend update and packaging improvements (including Debian license metadata); build/lockfile fixes for reproducible installs

Known issues

  • Pre-release software: agent, game, cloud, and local Home Agent features may still change — review network and security settings before enabling LAN access
  • The installer can fail due to environmental issues. See the Troubleshooting guide for help
  • Systems with discrete GPU and integrated GPU may need to disable iGPU when installing
  • Systems with more than one GPU may need to disable idle GPUs when inferencing
  • An interrupted download can block models from downloading properly. To fix this, go to the model location, delete the temporary file, and try again
  • The Colorizer workflow is limited to Xe2-supported GPUs (LNL and BMG)
  • A PC restart may be needed after installation for optimal functionality and performance
  • Reinstalling over an existing AI Playground installation will delete any ComfyUI custom nodes
  • Users can adjust settings beyond their specific system's ability, resulting in failed generations
  • Upgrading from 3.1.x may use a different shared model folder layout — verify model paths in Models settings after install

Version 3.2.0 beta supported hardware

  • Intel Core Series 3 Processors with 12GB+ of system memory (WCL)
  • Intel Core Ultra Series 3 Processors (PTL)
  • Intel Core Ultra Series 2 (H) Processors (ARL-H)
  • Intel Core Ultra Series 2 (V) Processors (LNL)
  • Intel Core Ultra Series 1 (H) Processors (MTL-H)
  • Intel Arc B Series GPU Cards (BMG)
  • Intel Arc A Series GPU Cards with 8GB+ Memory (ACM)
  • Nvidia GeForce RTX GPUs
  • Home Agent feature requires a minimum of 32GB RAM on laptops, or 16GB with a discrete GPU
Assets 4
Loading