Starstruk Logo MarkStarstruk

v1.8.0

FranckyB/Voice-Clone-Studiov1.8.0Feb 22, 2026by FranckyB

AI Summary

Major update introducing Prompt Hub for managing saved prompts across all generation tools, Ollama as an alternative LLM backend for local prompt generation, and Split by Paragraph feature for batch audio generation from multi-paragraph text.

Key Highlights

  • Prompt Hub integration with one-click transfer to tools
  • Ollama support as alternative LLM backend alongside llama.cpp
  • Split by Paragraph for batch audio generation
  • Category filtering shows only relevant prompt types
  • Combined preview plays all generated clips automatically

New Features

  • Prompt Hub with built-in Prompt Loader accordion in all generation tools
  • One-Click Transfer - send saved prompts directly to text input
  • Category Filtering - prompts organized by tool type
  • Auto-Refresh on accordion expand
  • Ollama LLM backend configurable in Settings
  • Use any local Ollama model for prompt generation
  • Split by Paragraph checkbox in Voice Clone
  • Base Name Prompt for sequential file naming
  • Combined Preview of all generated clips
  • Name Collision Check with client-side validation
  • Compatible with Qwen, VibeVoice, LuxTTS, and Chatterbox engines

Full Release Notes

#### Version 1.8.0 - Prompt Hub, Ollama Support & Split by Paragraph

**Prompt Hub Integration**
- **Prompts Directly in Tools** - Every generation tool (Voice Clone, Voice Presets, Voice Design, Conversation, Sound Effects) now has a built-in Prompt Loader accordion for quick access to saved prompts
- **One-Click Transfer** - Select a saved prompt and send it straight to the tool's text input without switching tabs
- **Category Filtering** - Prompts are organized by category, with each tool showing only the relevant prompt type
- **Auto-Refresh on Expand** - Prompt list refreshes automatically when the accordion is opened, no manual refresh needed

**Ollama Support**
- **New LLM Backend** - Added Ollama as an alternative LLM backend for prompt generation alongside llama.cpp
- **Configurable in Settings** - Choose between llama.cpp and Ollama in Preferences; Ollama connects to a locally running instance
- **Any Ollama Model** - Use any model available in your local Ollama installation for prompt generation

**Split by Paragraph**
- **Batch Audio Generation** - New "Split by Paragraph" checkbox in Voice Clone splits text by line breaks and generates a separate audio clip for each paragraph
- **Base Name Prompt** - Enter a base name for the batch; clips are numbered sequentially (e.g., MyScene_001, MyScene_002)
- **Combined Preview** - A combined preview of all clips plays automatically in the audio player after generation
- **Name Collision Check** - Client-side validation warns if files with the same base name already exist in the output folder
- **Works with All Engines** - Compatible with Qwen, VibeVoice, LuxTTS, and Chatterbox