StableStudio

StableStudio implements a standardized plugin interface (defined in Plugin.ts) that decouples the React-based UI from heterogeneous backend services, allowing seamless switching between Stability AI cloud APIs, local stable-diffusion-webui instances, or custom backends without UI changes. Each plugin implements methods for image creation, model/sampler retrieval, and authentication, enabling a provider-agnostic generation pipeline that routes requests through a unified interface layer.

Implements a text-to-image pipeline that accepts natural language prompts and routes them through the selected plugin backend (Stability AI API or local stable-diffusion-webui) with configurable generation parameters including model selection, sampler type, guidance scale, and seed. The Generation Module marshals the prompt into a backend-specific request format, handles async image streaming/polling, and returns rendered image buffers to the canvas component.

Provides a theming system that allows users to customize the application's visual appearance (colors, fonts, layout) through a centralized theme configuration, enabling light/dark mode support and custom branding. The Theme Module abstracts visual styling from component logic, enabling consistent theming across all UI components without duplicating style definitions.

Implements a request translation layer that converts UI parameters (prompt, model, sampler, guidance scale) into backend-specific API request formats, handling differences in parameter naming, value ranges, and request structure across Stability AI and stable-diffusion-webui APIs. This abstraction enables the UI to use consistent parameter names while supporting heterogeneous backends with different API contracts.

Provides an image editing pipeline that accepts an existing image and optional mask, applies AI-guided modifications through the selected backend's image-to-image capability, and renders the edited result back to the canvas. The Editor Module integrates with the canvas rendering system to support mask drawing, strength/guidance parameter adjustment, and real-time preview of inpainting results, enabling non-destructive iterative editing workflows.

Implements a capability discovery system where each plugin exposes available models and samplers through standardized methods (getModels(), getSamplers()), which the UI queries at initialization and caches for dropdown/selection components. This enables the UI to dynamically adapt to backend capabilities without hardcoding model lists, supporting backends with different model inventories and sampler implementations while maintaining a consistent selection interface.

Provides a configuration system for fine-grained generation control including guidance scale (classifier-free guidance strength), step count, seed, and sampler-specific parameters (e.g., scheduler type, noise schedule). The Advanced Settings component dynamically exposes sampler-specific controls based on the selected sampler type, marshaling these parameters into backend-specific request formats while maintaining a consistent parameter naming convention across providers.

Implements a canvas rendering system (likely using HTML5 Canvas or WebGL) that displays generated/edited images, manages layer composition for mask overlays and inpainting previews, handles zoom/pan interactions, and provides real-time feedback during generation. The Canvas component integrates with the Generation and Editor modules to display results, supports mask drawing for inpainting workflows, and manages the visual state of the application.

Provides a pre-built plugin that connects StableStudio to Stability AI's cloud image generation APIs, handling OAuth authentication flow, API credential management, and request/response marshaling. The plugin implements the standard Plugin interface, exposing Stability AI's available models and samplers, translating UI parameters into Stability AI API request formats, and handling rate limiting and error responses from the cloud service.

Provides a pre-built plugin that connects StableStudio to a local stable-diffusion-webui instance via HTTP API, enabling on-premise image generation without cloud dependencies. The plugin translates UI parameters into stable-diffusion-webui API request formats, handles long-polling for generation completion, and exposes the local instance's available models and samplers dynamically, supporting custom LoRA models and fine-tuned checkpoints.

Implements an environment configuration system that reads plugin settings (API keys, endpoints, model paths) from environment variables or configuration files, enabling secure credential management without hardcoding secrets in the application. The Environment Module abstracts backend-specific configuration details, allowing users to switch plugins by changing environment variables without modifying code or rebuilding the application.

Implements a React-based web interface that composes generation, editing, and configuration components into a cohesive application, managing application state through React hooks or context, and providing responsive layout that adapts to different screen sizes. The UI layer is decoupled from backend logic through the plugin system, enabling the same interface to work with different generation backends while maintaining consistent UX across providers.