Resolver is a high-performance, native Android interface for AI image generation. It is built using a Hybrid Architecture (Capacitor 6.0 + Vanilla JS) and utilizes Native Android Foreground Services to ensure generation queues remain active in the background or while the screen is off.
 Home / SDXL TAB |
 Flux UI |
 QWEN / Z-IMAGE TAB |
 Inpainting |
 Lora Tab |
 Configuration Tab |
 Magic Prompt Tab |
 Gallery Tab |
 Metadata viewer Tab |
 Queue Tab |
- Primary Integration: Resolver is natively optimized for WebUI Forge Neo (Haoming02) and the original lllyasviel Forge WebUI.
- Native API Communication: Utilizes direct API endpoints for rapid state synchronization and command execution.
- Smart Connection Mode: Toggle between Local Mode (automatic port appending) and External Mode (enforced HTTPS for Ngrok/Cloudflare).
- Remote Power Control: Integrated "Bojro Dev Power" for Wake-on-LAN and a remote KILL! signal to stop the backend server instantly.
- Workflow Orchestration: Execute API-compatible ComfyUI
.jsonworkflows directly within the app. - Dynamic UI Generation: Automatically parses workflow graphs to generate native sliders, dropdowns, and text inputs for samplers, checkpoints, and prompts.
- Live WebSocket Feedback: Real-time step monitoring and binary image previews via persistent socket connections.
- Unified Gallery: Automatically archives ComfyUI outputs into the main history for cross-engine metadata analysis.
- Performance Boost: Native support for the Forge Block Cache extension to reduce generation times.
- Optimization: Skips redundant diffusion steps for
bnb-nf4andfp8Flux models. - Direct Control: Threshold and skip step adjustments accessible directly via the Flux UI.
- Upscaling Engine: Full support for
Hires. fixacross SDXL, Flux, and Qwen modes. - Customization: Adjustable upscalers (ESRGAN, etc.), denoising strength, and upscale factors.
- Foreground Service: Java-based service with Wake Locks to maintain active WebSocket connections during long Flux GGUF tasks or batch queues.
- Persistence: Robust management for Ongoing, Next, and Completed jobs.
- Flux GGUF: Support for VAE, CLIP, and T5-XXL selectors with specific quantization bit settings.
- Qwen / Z-Image Turbo: Specialized mode for Qwen models with dense narrative support.
- SDXL Powerhouse: Comprehensive control over sampling, scheduling, and aspect ratio locking.
- Mobile Inpainting: Touch-optimized canvas editor with mask blurring and denoising controls.
- Smart Expansion: Connects to local LLM servers (LM Studio, Ollama) to translate simple ideas into complex prompts.
- Mode Awareness: Automatically switches system instructions based on the active engine (SDXL, Flux, or Qwen).
WebUI Forge:
In webui-user.bat, set:
set COMMANDLINE_ARGS=--listen --api --cors-allow-origins *
ComfyUI:
Append these flags to your launch command:
--listen --enable-cors-header *
Host a Bojro PromptMaster model on your PC:
- Download: PromptMaster v2 (Uncensored) or v1.
- Server: Load into LM Studio or Ollama on your local network.
- Connect: Open the Bot modal in Resolver, enter the server URL, and save.
Requires BojroPowerv_x_portable.exe on the host PC.
- Run Helper: Execute the utility on your PC.
- Configure: Set the Wake Port (Default:
5000) in the CFG Tab. - Operation: Use the Power Button (ტ) to start or the KILL! button to halt services.
- Install WebUI Forge and/or ComfyUI on your PC.
- Download the latest
.apkfrom the Releases. - Ensure phone and PC are on the same Wi-Fi.
- Set PC IP in the CFG Tab and save.
git clone [https://github.com/bojrodev/Resolver-WebUI-Forge-Client.git](https://github.com/bojrodev/Resolver-WebUI-Forge-Client.git)
npm install
npx cap sync
npx cap open android