Draw Things

Executes Stable Diffusion and FLUX models directly on Apple Silicon devices using Metal GPU acceleration, downloading models to local storage and performing inference without cloud transmission. The architecture leverages Metal's compute shaders for parallel tensor operations, enabling real-time generation on M-series chips while maintaining complete data privacy for prompts and generated images in the free tier.

Supports Low-Rank Adaptation (LoRA) training directly on Apple Silicon devices, allowing users to fine-tune base models (Stable Diffusion, FLUX) with custom datasets without cloud infrastructure. The implementation uses LoRA's parameter-efficient approach (adapting only low-rank matrices rather than full model weights) to reduce memory footprint and training time, with trained LoRAs stored locally and optionally uploaded to Draw Things+ cloud for inference.

Provides programmatic access to Draw Things' inference capabilities (local or cloud) for integration into third-party applications, enabling developers to embed image generation into their own tools. The implementation exposes an API (specification unspecified) with authentication and supports both local device inference and cloud compute, though exact endpoint structure, authentication mechanism, and SDK availability are undocumented.

Generates multiple images in sequence with varying parameters (different prompts, seeds, guidance scales, or models) to explore design space efficiently. The implementation queues generation tasks and executes them sequentially on local hardware or cloud infrastructure, allowing users to specify parameter ranges or lists and receive multiple outputs.

Provides UI controls and presets for fine-tuning generation parameters (guidance scale, sampling steps, seed, sampler algorithm, negative prompts) to control output quality, style, and consistency. The implementation exposes these parameters through sliders, text inputs, and preset templates, allowing users to iteratively refine generation without code.

Enables targeted image modification by accepting a base image, mask, and text prompt, then regenerating only the masked region using the diffusion model while preserving unmasked areas. The implementation uses latent-space inpainting (encoding the image to latent space, masking the latent representation, and diffusing only masked regions) to maintain coherence with surrounding content while applying new generation semantics from the prompt.

Extends image boundaries in any direction (up, down, left, right, or arbitrary angles) by generating new content that seamlessly blends with existing edges. The implementation uses outpainting (a variant of inpainting where the model generates content outside the original image bounds) combined with edge-aware context blending to maintain visual continuity and perspective consistency across the expanded canvas.

Integrates ControlNet (a neural network adapter that conditions diffusion models on structural inputs like edge maps, depth maps, pose skeletons, or semantic segmentation) to guide image generation toward specific compositions, layouts, or structural constraints. The implementation loads ControlNet weights alongside base models and uses multi-scale feature injection to influence generation while maintaining semantic fidelity to text prompts.

Converts static images into short video clips by generating frame sequences that maintain visual consistency and smooth motion. The implementation approach is unspecified in documentation (could be frame interpolation, latent-space video diffusion, or optical flow-based synthesis), but the capability enables animation of still images with semantic understanding of motion from text prompts.

Applies artistic styles to images (e.g., photo-to-animation, realistic-to-cartoon, style-specific rendering) by using diffusion models conditioned on style descriptors or reference images. The implementation leverages text prompts with style keywords and optional reference images to guide the model toward specific artistic outputs while preserving content structure.

Generates images of clothing, accessories, or character designs applied to reference images (e.g., 'see how this shirt looks on a model' or 'visualize this character design on different body types'). The implementation uses conditional image generation with spatial awareness to place and adapt designs onto reference images while maintaining realistic lighting, proportions, and integration.

Provides optional cloud inference via Draw Things' managed servers for users on Community and Draw Things+ tiers, allowing users to offload generation to cloud infrastructure when local device is unavailable or when higher-speed inference is desired. The architecture maintains local-first default (free tier) while enabling seamless cloud fallback through account-based tier detection and API routing.

Manages model lifecycle (discovery, download, storage, versioning) through Draw Things' proprietary model server, allowing users to browse available models (Stable Diffusion, FLUX, and community models), download them to local storage, and switch between versions. The implementation caches downloaded models locally to avoid re-downloading and provides UI for model selection and management.