The libprocessing roadmap is split into a few high-level categories, plus a foundation that everything else depends on:

0. Foundations: stability, docs, and distribution that make everything else trustworthy.
1. Core rendering technologies: exposing engine-level features in Processing's core API.
2. Bringing historical 3rd-party library functionality into Processing itself: e.g. audio, MIDI, video, etc.
3. Narrow features for professionals, installation artists, etc.: e.g. NDI support, camera SDKs, lasers.
4. Advanced rendering techniques: rendering beyond the triangle rasterizer, e.g. SDFs, volumetrics, ray tracing.

Platform targets (desktop, web/WASM, embedded, XR) cut across all of these.

• Stable release on upstream bevy, published to crates.io
• CI with example, visual, and performance regression testing
• Documentation site with architecture, contributing guide, API reference, and tutorials
• Language bindings beyond Rust (Python via mewnala, Java for Processing, JS for p5.js)
• Live coding and hot reload of sketches and shaders

• Full Processing API coverage
• Compute shaders and full-screen post-processing effects
• GPU particles and simulation, like the 3D flocking demo from SIGGRAPH (#188, done)
• HDR pipeline with EXR read/write (#178)
• PBR materials and lighting
• Shadows, bloom, depth of field, and other engine-level effects
• Modern custom shader support (WGSL)
• GPU instancing for drawing very large numbers of shapes
• High-quality text rendering and font outlines
• GUI built on bevy UI
• Web/WASM parity, including p5.js as a libprocessing consumer

• Audio playback, effects, routing, and analysis via bevy_seedling (replaces Sound/Minim)
• MIDI (replaces The MidiBus)
• Video playback, webcam capture, and video/GIF export (replaces processing-video, VideoExport, gifAnimation)
• OSC and general networking: UDP/TCP, WebSockets, HTTP (replaces oscP5/netP5)
• Serial I/O
• Orbit/easy 3D camera controls (replaces PeasyCam)
• 2D and 3D physics (replaces Box2D for Processing, Fisica, toxiclibs physics)
• Computational geometry and mesh tools: meshing, booleans, subdivision, Voronoi (replaces toxiclibs, HE_Mesh)
• Tweening and animation easing (replaces Ani)
• Computer vision: blob detection, optical flow, face/pose tracking (replaces OpenCV for Processing, BlobDetection)
• Machine learning inference for vision, pose, and generative models
• Depth sensors and hand tracking (replaces SimpleOpenNI, Kinect and Leap Motion libraries)
• 3D model import: glTF and FBX
• PDF/SVG vector import and export, including plotter output (replaces the PDF/SVG libraries, Geomerative)
• Maps and geospatial data (replaces Unfolding)

• Show control: DMX, Art-Net/sACN, CAN
• Lasers: ILDA DAC output (Ether Dream, Helios, LaserCube) and ILDA file support
• LED pixel mapping and controllers (e.g. Open Pixel Control, WLED)
• Inter-app video: NDI, Syphon, Spout
• Capture and output cards (Blackmagic DeckLink, AJA)
• Depth cameras and machine-vision camera SDKs
• Motion-capture and tracking systems (OptiTrack, Vicon, VRPN)
• Time sync: LTC/MIDI timecode, Ableton Link, multi-machine frame sync
• IoT and control messaging: MQTT
• Embedded: an Arduino-like interface for microcontrollers (ESP32, RP235x, etc.)
• Multi-display, projection mapping, and edge blending
• XR via OpenXR
• Pressure-sensitive pen tablet input
• Pipeline interchange: USD and Alembic
• Installation-grade runtime: headless/kiosk mode and long-running stability

• SDF rendering and ray marching as first-class drawing primitives
• Volumetrics: fog, clouds, volume textures, OpenVDB
• Gaussian splatting via PLY (#207)
• Point cloud rendering at scale
• Real-time ray tracing and global illumination
• Path tracing for offline, high-quality stills
• GPU fluid, cloth, and reaction-diffusion simulation
• Procedural and compute-driven geometry
• Non-photorealistic rendering: toon shading, outlines, hatching, sketchy styles