Awesome Creative Intelligence

A curated list of tools, frameworks, and codebases for creative intelligence in code.

Creative intelligence sits at the intersection of computation, art, and artificial intelligence. This list catalogs the open-source projects, platforms, and research codebases that enable machines and humans to create together -- spanning visual art, music, writing, architecture, science, and beyond.

---

Contents

• Creative Coding Frameworks
• AI Image Generation
• AI Video Generation
• AI Music and Audio
• Live Coding
• 3D Generation
• Evolutionary and Genetic Art
• Procedural Generation
• Creative Writing AI
• Neural Style Transfer
• Creative Agent Systems
• Scientific Creative AI
• Design and Architecture
• Generative Art Platforms
• Curated Lists
• Contributing

---

Creative Coding Frameworks

Libraries and environments for writing code that produces visual, sonic, or interactive art.

• p5.js — A JavaScript library for creative coding, making coding accessible for artists, designers, educators, and beginners. The spiritual successor to Processing for the web. JavaScript ~22k stars

• Processing — The original creative coding environment that launched a movement. A flexible software sketchbook and language for learning how to code within the context of the visual arts. Java

• openFrameworks — An open-source C++ toolkit designed for creative coding. Wraps together several commonly used libraries for graphics, audio, video, and interaction into a consistent, cross-platform interface. C++ ~10k stars

• TouchDesigner — A node-based visual programming environment for real-time interactive multimedia content. Used extensively in live performances, installations, and projection mapping. Visual/Python

• Nannou — A creative coding framework for Rust, offering an expressive API for generative art, audio visualization, and interactive installations with the safety and performance of Rust. Rust ~4k stars

• three.js — The dominant JavaScript 3D library. Powers countless creative web experiences, from data visualizations to immersive art pieces running in the browser with WebGL and WebGPU. JavaScript ~103k stars

• canvas-sketch — A loose collection of tools, modules, and resources for creating generative art in JavaScript and the browser. Includes utilities for exporting high-resolution prints. JavaScript ~5k stars

• Hydra — A platform for live coding networked visuals inspired by analog video synthesizers. Runs in the browser with a minimal, composable syntax for real-time visual performance. JavaScript ~2k stars

• Shadertoy — A community and platform for creating and sharing fragment shaders. Home to some of the most impressive real-time rendering experiments on the web. GLSL

• Context Free Art — A program that generates images from written instructions called grammars. Context Free uses a variation of context-free grammars to produce stochastic, recursive visual art. C++

• ml5.js — Friendly machine learning for the web, built on top of TensorFlow.js. Provides immediate access to pre-trained models for image classification, pose detection, sound recognition, and generative text within creative coding projects. JavaScript ~6k stars

• Pts.js — A library for visualization and creative coding. Pts enables you to compose and express what you see in code, bridging the gap between algorithmic thinking and visual expression. TypeScript

AI Image Generation

Models and tools for generating, editing, and composing images with AI.

• FLUX.1/FLUX.2 — State-of-the-art open-weight text-to-image models from Black Forest Labs (the team behind Stable Diffusion). FLUX.1 comes in schnell (fast), dev, and pro variants; FLUX.2 pushes fidelity and controllability further. Python

• Stable Diffusion XL — The influential open-weight image generation model from Stability AI. SDXL introduced a base + refiner architecture and 1024x1024 default resolution, becoming the foundation for an enormous ecosystem of fine-tunes and tools. Python

• ComfyUI — The most powerful and modular node-based GUI and backend for diffusion models. Supports Stable Diffusion, FLUX, video models, and virtually any diffusion pipeline through a visual graph interface. Python ~60k stars

• Fooocus — An image generating software inspired by the simplicity of Midjourney. Rethinks Stable Diffusion and FLUX workflows by automating complex prompt engineering and pipeline tuning behind a minimal interface. Python ~42k stars

• HuggingFace Diffusers — The go-to library for state-of-the-art pretrained diffusion models. Provides a unified API for image, audio, and 3D generation with schedulers, pipelines, and a massive model hub. Python ~26k stars

• CLIP / OpenCLIP — Open-source implementation of OpenAI's Contrastive Language-Image Pre-training. The backbone of text-guided image generation, enabling models to understand the relationship between text descriptions and visual content. Python

• DALL-E 2 PyTorch — An independent PyTorch implementation of DALL-E 2 (unCLIP), OpenAI's text-to-image system using CLIP latents, a prior network, and a diffusion decoder. A key reference implementation for the research community. Python ~11k stars

AI Video Generation

Models and pipelines for generating video from text, images, or other video.

• Open-Sora — An open-source initiative to efficiently reproduce OpenAI's Sora video generation model. Implements a diffusion transformer architecture for high-quality text-to-video and image-to-video generation. Python ~23k stars

• CogVideoX — A large-scale text-to-video generation model from Tsinghua University's Knowledge Engineering Group. Leverages a 3D variational autoencoder and expert transformer blocks for coherent, temporally consistent video. Python ~10k stars

• HunyuanVideo — Tencent's 13-billion-parameter open video generation model. One of the largest open-source video generation models, capable of producing high-resolution, temporally coherent video from text descriptions. Python 13B params

• Wan 2.1/2.2 — Alibaba's suite of open video generation models supporting text-to-video, image-to-video, and video editing. Wan 2.2 adds improved motion quality and longer generation capabilities. Python

• Mochi 1 — A high-quality open-source video generation model from Genmo. Focuses on motion quality and temporal coherence, producing smooth, natural-looking video from text prompts. Python

• AnimateDiff — A practical framework for animating personalized text-to-image diffusion models without requiring model-specific tuning. Inserts a motion module into frozen text-to-image models to produce animated outputs. Python

AI Music and Audio

Models and tools for generating music, speech, sound effects, and audio with AI.

• Bark — A transformer-based text-to-audio model from Suno that can generate realistic multilingual speech, music, sound effects, and nonverbal cues like laughter and sighing from text prompts. Python ~39k stars

• AudioCraft / MusicGen — Meta's library for audio generation research, containing MusicGen (text/melody-to-music), AudioGen (text-to-sound), and EnCodec (neural audio codec). A unified codebase for state-of-the-art audio AI. Python ~21k stars

• Magenta — Google's research project exploring the role of machine learning in creative processes. Includes models for melody generation, music transcription, artistic image stylization, and more. Python/JavaScript ~19k stars

• Jukebox — OpenAI's generative model for music, producing raw audio in a variety of genres and artist styles. Generates minute-plus samples with singing, using a multi-scale VQ-VAE approach. Python ~8k stars

• YuE — A music generation model capable of producing full songs with vocals, lyrics, and instrumental accompaniment. Pushes toward end-to-end song creation from text descriptions. Python

• ACE-Step — A step-distilled music generation model enabling fast, high-quality music creation. Reduces inference time dramatically while maintaining the quality of full diffusion sampling. Python

• DiffRhythm — A diffusion-based model for generating rhythmic music. Focuses on producing music with strong rhythmic structure and temporal coherence using diffusion processes. Python

• Riffusion — A creative approach to music generation that produces music by generating spectrogram images with Stable Diffusion and converting them back to audio. An elegant demonstration of cross-modal generation. Python

Live Coding

Environments for writing code that generates music or visuals in real time, often in performance settings.

• Sonic Pi — A code-based music creation and performance tool. Designed for education and live performance, it turns coding into a musical instrument with an emphasis on accessibility and expressiveness. Ruby/C++ ~11k stars

• SuperCollider — A platform for audio synthesis and algorithmic composition, used by musicians, artists, and researchers. Provides a powerful interpreted language (sclang) and real-time audio server (scsynth). C++ ~5k stars

• TidalCycles — A language for live coding musical patterns. Tidal uses a Haskell-based pattern language to describe flexible, polyphonic, polyrhythmic sequences that can be transformed in real time. Haskell ~2k stars

• Strudel — A JavaScript port and reimagining of TidalCycles that runs entirely in the browser. Makes algorithmic music accessible without any installation, with a live coding editor and instant audio feedback. JavaScript ~2k stars

3D Generation

Models and tools for generating 3D objects, scenes, and representations from text or images.

• 3D Gaussian Splatting — The original implementation of 3D Gaussian Splatting for real-time radiance field rendering. Represents scenes as collections of 3D Gaussians for fast, high-quality novel view synthesis. Python/CUDA ~15k stars

• Shap-E — OpenAI's model for generating 3D objects conditioned on text or images. Produces implicit 3D representations that can be rendered as both textured meshes and neural radiance fields. Python ~11k stars

• Nerfstudio — A modular framework for neural radiance field development. Provides a unified pipeline for training, visualizing, and exporting NeRFs with a plug-and-play architecture for different NeRF methods. Python ~10k stars

• Point-E — OpenAI's system for generating 3D point clouds from text prompts. Uses a text-to-image diffusion model followed by an image-to-3D model, trading some quality for dramatically faster generation speed. Python ~6k stars

Evolutionary and Genetic Art

Algorithms that evolve visual art, designs, and solutions using principles from biological evolution.

• DEAP — Distributed Evolutionary Algorithms in Python. A comprehensive evolutionary computation framework supporting genetic algorithms, genetic programming, evolution strategies, and multi-objective optimization. Python ~5.7k stars

• neat-python — A Python implementation of the NEAT (NeuroEvolution of Augmenting Topologies) algorithm. Evolves neural network topology and weights simultaneously, used in creative applications from game AI to generative art. Python ~1.5k stars

• EvoJAX — A hardware-accelerated neuroevolution toolkit from Google built on JAX. Enables massively parallel evolutionary strategies on GPUs/TPUs for evolving neural network controllers, policies, and creative agents. Python ~800 stars

• pyribs — A bare-bones Python library for quality-diversity optimization. Implements MAP-Elites and related algorithms that produce diverse, high-performing solutions -- ideal for generating varied creative outputs. Python ~400 stars

• QDax — A JAX-based library for quality-diversity algorithms and neuroevolution. Provides hardware-accelerated implementations of MAP-Elites variants for exploring diverse solution spaces at scale. Python

• GeneticPainter — An evolutionary algorithm that approximates images using primitives like polygons, circles, and brush strokes. Evolves populations of candidate paintings through selection and mutation toward a target image. Python

• CGP-Evolutionary-Art — Cartesian Genetic Programming applied to evolutionary art. Evolves mathematical expressions that map pixel coordinates to colors, producing abstract and organic visual patterns. Python

Procedural Generation

Algorithms and systems for creating content through rules, constraints, and stochastic processes.

• WaveFunctionCollapse — A constraint-solving algorithm inspired by quantum mechanics that generates tile maps and textures from small example images. Produces locally similar but globally novel outputs through iterative constraint propagation. C# ~23k stars

• Material Maker — A procedural materials authoring tool built with Godot. Provides a node-based interface for creating PBR materials, textures, and patterns using mathematical functions and noise generators. GDScript

• OpenAI Procgen — A suite of 16 procedurally generated game environments designed for reinforcement learning research. Each environment produces an infinite variety of levels, testing generalization in visually rich settings. C++/Python

Creative Writing AI

Tools for collaborative fiction, interactive narratives, and AI-assisted creative writing.

• KoboldAI — A browser-based front-end for AI-assisted writing with local and remote language models. Supports story mode, adventure mode, and chat mode with extensive model support and memory management. Python

• SillyTavern — A user interface for interacting with language models for creative writing and roleplay. Provides character management, world-building tools, and extensive customization for narrative AI experiences. JavaScript

• StoryCraftr — An AI-powered toolkit for writing books, screenplays, and long-form narratives. Provides structure-aware generation that respects plot arcs, character consistency, and narrative conventions. Python

Neural Style Transfer

Techniques for transferring artistic styles between images and generating art through neural network optimization.

• DeepDream — Google's original neural network visualization technique that became an art movement. Amplifies patterns detected by a neural network through gradient ascent, producing psychedelic, fractal-like imagery. Python ~13k stars

• VQGAN+CLIP — The combination that launched the text-to-image art revolution before diffusion models. Optimizes a VQGAN image generator using CLIP's text-image understanding, producing surreal, dreamlike art from text prompts. Python

• CAN / AICAN — Creative Adversarial Networks that generate art by maximizing deviation from established styles while remaining within the distribution of art. Trained to produce novel aesthetics rather than imitate existing ones. Python

Creative Agent Systems

Autonomous and semi-autonomous AI systems that can plan, execute, and iterate on creative tasks.

• AutoGPT — A pioneering autonomous AI agent that chains LLM calls with tool use to accomplish complex tasks. Demonstrated the potential of self-directed AI systems for creative and technical workflows. Python ~173k stars

• LangChain — A framework for developing applications powered by language models. Provides composable chains, agents, memory, and tool integrations that enable complex creative AI workflows and multi-step reasoning. Python ~100k stars

• CrewAI — A framework for orchestrating role-playing, autonomous AI agents. Agents collaborate with defined roles, goals, and backstories to tackle complex creative and analytical tasks as a team. Python ~26k stars

• AIDM — AI Dungeon Master, a system for running tabletop roleplaying games with AI-driven narrative generation. Manages world state, NPC behavior, and narrative coherence for interactive storytelling. Python

• GameMasterAI — An AI-powered game master that generates dynamic narratives, encounters, and world events for tabletop and digital RPGs using language model reasoning. Python

• K:BOT — An open-source terminal AI agent with 37 specialist agents including creative, writing, design, and analysis capabilities. Features autonomous planning, learning from interactions, and multi-provider AI support. TypeScript GitHub

Scientific Creative AI

AI systems that exhibit creative intelligence in scientific discovery, protein design, and mathematical reasoning.

• AlphaFold — DeepMind's protein structure prediction system that solved a 50-year grand challenge in biology. Predicts 3D protein structures from amino acid sequences with atomic accuracy, enabling creative protein engineering. Python ~13k stars

• AlphaEvolve — DeepMind's Gemini-powered coding agent that designs novel algorithms through evolutionary search. Discovered new mathematical constructs and optimized real-world systems by creatively exploring solution spaces. Python

• AI Co-Scientist — Google's multi-agent AI system designed to function as a virtual scientific collaborator. Generates novel hypotheses, designs experiments, and synthesizes research across disciplines through creative scientific reasoning. Python

Design and Architecture

Tools for computational design, parametric architecture, color theory, and AI-assisted design workflows.

• Grasshopper 3D — A visual programming language and environment for computational design within Rhino 3D. The standard tool for parametric architecture, enabling designers to create complex forms through algorithmic rules. C#

• Figma Make — Figma's AI-powered design generation feature that creates UI designs and components from natural language descriptions. Integrates AI directly into the professional design workflow. Web

• Colormind — A color scheme generator powered by deep learning. Trained on photographs, movies, and popular art, it generates aesthetically coherent color palettes that capture the mood and style of its training data. Web

• Khroma — An AI color tool that learns your color preferences and generates personalized palettes. Uses machine learning to understand individual aesthetic taste and suggest harmonious combinations. Web

Generative Art Platforms

Platforms and marketplaces for creating, collecting, and experiencing generative art.

• fxhash — An open platform for generative art on the Tezos blockchain. Artists deploy generative algorithms, and each mint produces a unique output from the code, embracing controlled randomness as artistic medium. Web

• Art Blocks — A curated platform for on-chain generative art on Ethereum. Artists store their generative algorithms on the blockchain, where each token generates a unique artwork at the moment of minting. Web

• Artbreeder — A collaborative art tool that uses generative adversarial networks (GANs) to blend and evolve images. Users can combine multiple images, adjust attributes, and explore latent space to discover new visual forms. Web

Curated Lists

Meta-lists and collections that catalog creative intelligence resources.

• awesome-creative-coding — A comprehensive curated list of creative coding resources including books, courses, tools, frameworks, and communities. The definitive starting point for the creative coding ecosystem. ~13k stars

• awesome-generative-ai — A curated list of modern generative AI projects and services covering image, video, audio, text, and code generation. Maintained with regular updates as the field evolves. ~6k stars

• the-gan-zoo — A comprehensive catalog of every named GAN (Generative Adversarial Network) variant published in research. An essential taxonomy of the GAN family tree, from the original 2014 paper through hundreds of variants. ~15k stars

---

Contributing

Contributions are welcome. This list is maintained by the kernel.chat group community.

To add a project:

1. Fork the repository
2. Add your entry to the appropriate category following the format: **[Name](url)** -- Description. \Language` `~Nk stars``
3. Keep entries sorted by star count (descending) within each category
4. Submit a pull request

Guidelines

• Projects should be actively maintained or historically significant
• Include open-source projects, notable closed-source tools, and research codebases
• Descriptions should be concise (1-2 sentences) and focus on what makes the project notable
• Star counts are approximate and should be updated periodically

---

License

To the extent possible under law, the contributors have waived all copyright and related or neighboring rights to this work.