This project provides an Android client application for Stable Diffusion AI, enabling users to generate digital art on their mobile devices. It targets users who want a flexible and privacy-focused way to interact with various Stable Diffusion backends, offering a no-ads, no-telemetry experience.

How It Works

The app acts as a front-end, connecting to different Stable Diffusion backends via their APIs. It supports a wide array of providers, including self-hosted Stable-Diffusion-WebUI (AUTOMATIC1111), SwarmUI, AI Horde, Hugging Face Inference API, OpenAI (DALL-E 2/3), and Stability AI. It also offers experimental local generation capabilities using Microsoft ONNX Runtime and Google AI MediaPipe on the device.

Quick Start & Requirements

• Installation: Download from Google Play, F-Droid, or GitHub releases.
• Backend Setup: Requires configuring and running a compatible Stable Diffusion backend (e.g., AUTOMATIC1111 with --api --listen flags) or obtaining API keys for cloud services.
• Resources: Local generation modes (Beta) utilize device resources. Remote backends require network connectivity and a running server or API access.
• Documentation: Build flavor difference

Highlighted Details

• Supports Txt2Img and Img2Img modes with dynamic resolution up to 2048px.
• Offers extensive control over sampling methods, steps, CFG scale, and image restoration.
• Includes advanced features for AUTOMATIC1111 backends like Lora, Textual Inversion, and Hypernetwork pickers.
• Features an in-app gallery with zoom, pinch, and generation info display, with export options.

Maintenance & Community

The project is actively maintained by ShiftHackZ. Community contributions for translations are welcomed.

Licensing & Compatibility

The project is open-source, provided with no warranty. Specific licensing details are not explicitly stated in the README, but the "open source" nature suggests permissive usage.

Limitations & Caveats

Local generation modes are in Beta and may have limited functionality (e.g., only Txt2Img). Compatibility with specific backend versions or configurations is not guaranteed. The README does not detail specific hardware requirements for local generation.