画像
ai-media-generator
Claude Code skill for cinematic AI image, video and music prompts — 14+ platforms
★ 255 スターMIT
ガイドを見る →sprite-genは、aldegadが公開しているオープンソースの画像プロジェクトです（GitHub: aldegad/sprite-gen）。GitHubスターは2,616、フォークは275で、Apache-2.0ライセンスの下で公開されています。GitHubでの最終更新は2026年10月8日です。
Generate clean 2D game sprites and animation atlases from one drawing — a Codex/Claude skill and Python CLI
Prepare one base image of your character and the list of motion states you need (idle, walk, run, jump, attack...), and choose the atlas pipeline or the video-to-loop pipeline.
Clone the repo and install it into a fresh virtualenv, or install it as a skill into Codex/Claude.
Activate the venv, check sprite-gen --help, and connect a generation provider (codex or grok on your existing subscription; the video path uses your own grok login or XAI_API_KEY, plus ffmpeg/img2webp).
sprite-gen prepare your run, gen-set the state rows, extract the frames, and compose-atlas into sprite-sheet-alpha.png + manifest.json.
prepare → gen/gen-set → extract → compose-atlas produces a transparent atlas plus a manifest with absolute frame rectangles, per-state fps and loop flags — engines never guess a grid.
Chroma backgrounds are unmixed to true transparency with no leftover fringe, verified against white backgrounds.
A still idle pose becomes a living loop with deterministic, anatomy-aware squash-and-stretch baked from a sidecar field.
One still plus a video model yields seamless transparent GIF/WebP/strip loops cut at the clip's true period, with canvas sized per motion (tall for jumps, wide for attacks).
Recolor into deterministic colorways, compose rig layers, and export for Aseprite / Phaser / Flame — without regenerating.
Standalone cutout, slice-sheet and unpack-atlas utilities, background tiles, projected shadows, motion inspection, and an optional scene renderer for finished assets.
Clone sprite-gen and install it into a fresh virtualenv as the quickstart shows (the venv is the only supported interpreter).
git clone https://github.com/aldegad/sprite-gen.git
cd sprite-genPrint the command map, grouped by pipeline and tool domain.
sprite-gen --helpPrepare a run from one base image, generate the state rows, extract frames and compose the atlas.
sprite-gen prepare --out-dir <dir> --character-id <id> --base-image base.png
sprite-gen gen-set --run-dir <dir> --provider codex
sprite-gen extract --run-dir <dir>
sprite-gen compose-atlas --run-dir <dir>aldegad maintains it; the component-row workflow was inspired by the Apache-2.0 hatch-pet skill, per the README's attribution.
The tool is Apache-2.0 licensed. Generation uses providers on subscriptions you already pay for (codex/grok), an explicit-only OpenAI provider that bills per call, and your own xAI credentials for video.
Image generation rides on your existing Codex or Grok subscription by default; the video pipeline needs your own grok login or XAI_API_KEY. Nothing is shipped with the repo.
Raw generations drift off-grid, change faces between frames and won't key out cleanly. sprite-gen's pipeline locks identity, extracts real alpha frames and emits an atlas + manifest an engine can consume.
Clone, create a venv, pip install -e ., then run the prepare → gen-set → extract → compose-atlas flow on one base image.
Claude Code skill for cinematic AI image, video and music prompts — 14+ platforms
PhotoCraft — a clean-room Photoshop reimplementation in pure Rust: layers, masks, adjustment layers, type, vectors, brushes and real PSD files, offline, MIT / Apache-2.0
Extract clean line sketches from anime art, illustrations, and manga with a pretrained GAN model.
Claude Design: AI UI/UX architect.
DeepSeek's Janus series - unified multimodal models that both understand and generate images
Turn ComfyUI node workflows into runnable Python scripts you can automate and rerun anywhere.
LibreAIカタログに掲載された画像プロジェクトの中で最もスターの多いもの — sprite-genが合わなかった場合に検討したい、現実的な候補リストです。
Stable Diffusion web UI
The node-graph powerhouse for precise, professional AI image and video generation.
High performance self-hosted photo and video management solution
PhotoCraft — a clean-room Photoshop reimplementation in pure Rust: layers, masks, adjustment layers, type, vectors, brushes and real PSD files, offline, MIT / Apache-2.0
Focus on prompting and generating
カタログの中でsprite-genとスター数が近い画像プロジェクト — 前後の棚もそのまま見て回れます。
InfiniteYou (InfU) — flexible photo recrafting on FLUX that preserves your identity (ICCV 2025 Highlight)
CUT — contrastive unpaired image-to-image translation in PyTorch, faster and lighter than CycleGAN (ECCV 2020)
FLUX, Stable Diffusion, SDXL, SD3, LoRA, Fine Tuning, DreamBooth, Training, Automatic1111, Forge WebUI, SwarmUI, DeepFake, TTS, Animation, Text To Video, Tut…
Turn ComfyUI node workflows into runnable Python scripts you can automate and rerun anywhere.
Booru style tag autocompletion for AUTOMATIC1111's Stable Diffusion web UI
Extract clean line sketches from anime art, illustrations, and manga with a pretrained GAN model.