V0.12

[jamesk9526]

What’s New
MiniMax H3 Encoding

• Added optional qwen3vl_32b_minimax_h3_int8_convrot.safetensors text encoder.
• Fast NVFP4-AWQ remains the default.
• New “Slower · better encoding” quality option is available in Create → Advanced controls and Settings → Generation defaults.
• Quality mode will not silently fall back if the INT8 encoder is missing.
  Prompt Builder
• Added automatic reference summaries with visual source chips.
• Reference assignments can be reviewed in a dedicated modal.
• Added copyable prompt output and improved prompt insertion tools.
• Ollama prompt assistance now produces concise, timeline-aware additions that avoid repeating existing instructions.
  Video Generation
• Added MiniMax H3 Turbo Ref2VA 8-Step v1.0 support.
• Added exact continuation workflow using the previous render’s final frame.
• Continuations preserve canvas size, sampling settings, motion settings, and LoRA configuration.
• Removed the redundant manual final-frame extraction flow for continuation work.
  Rendering and Progress
• Added live video render progress for LTX 2.5 location-set generation.
• Progress now includes current stage, sampler steps, and percentage completion.
• Improved queue, cancellation, and render status feedback.
  Chat Workspace
• Enlarged the AI chat workspace.
• Added pop-out chat windows.
• Added always-on-top support.
• Added dock/restore controls for returning the chat to the main window.
• support for lmstudio as a llm provider
  Reference and Location Workflows
• Improved character, wardrobe, location, and reference-media assignment.
• Added clearer locked reference instructions with editable source management through modal controls.
• Improved responsive layouts for desktop, resized windows, and mobile views.
  Reliability
• Added persistence for encoder preferences and generation defaults.
• Expanded model detection and validation.
• Added regression coverage for official H3 model selections and Turbo workflows.
• Verified with linting, TypeScript checks, production build, and workflow tests.

---

This discussion was created from the release V0.12.