touchdesigner-mcp-server: Agent for Claude Code

You are researching and documenting cutting-edge experimental TouchDesigner techniques for the MCP server knowledge base.

Project location: /home/robert/Documents/TD-MCP/touchdesigner-mcp-server

First, read these existing files for context:

1. /wiki/data/python-api/glslTOP.json or /wiki/docs/python/glslTOP_Class.htm — GLSL TOP API
2. Any script/compute TOP references in the project
3. /wiki/data/tutorials/ directory — see what tutorials exist

Then create /wiki/data/experimental/ directory with these comprehensive data files:

a) /wiki/data/experimental/index.json: { "categories": ["glsl", "gpu-compute", "machine-learning", "generative-systems", "audio-visual", "networking", "python-advanced"], "totalTechniques": 0, "lastUpdated": "2026-02-21" }

b) /wiki/data/experimental/glsl.json — GLSL techniques: At minimum include these techniques with full detail:

• Raymarching / SDF rendering in GLSL TOP
• Reaction-diffusion systems via GLSL
• Feedback loop patterns (GLSL + Feedback TOP)
• Custom post-processing pipelines
• Procedural texture generation For each technique provide: name, description, operators[], codeExample (real GLSL code), performance (low/medium/high), versionRequirement, useCases[], relatedTechniques[]

c) /wiki/data/experimental/gpu-compute.json — GPU compute patterns:

• Script TOP for GPU buffer operations
• CUDA integration patterns
• Shared memory TOP/CHOP patterns
• GPU instancing via Geometry COMP
• Particle systems using compute shaders

d) /wiki/data/experimental/machine-learning.json — ML integration:

• TouchEngine CHOP/TOP for ML models
• ONNX model integration
• Stable Diffusion via Script TOP
• Real-time pose estimation (MediaPipe patterns)
• Audio ML (classification, generation)
• Version requirement: 2022+ for TouchEngine

e) /wiki/data/experimental/generative-systems.json — generative art techniques:

• L-system implementations in SOPs
• Cellular automata via DAT + Script
• Strange attractors (Lorenz, Rössler) in CHOPs
• Agent-based systems using Replicator COMP
• Recursive network patterns

f) /wiki/data/experimental/audio-visual.json — audio-visual synthesis:

• FFT analysis to geometry (CHOP to SOP)
• Granular synthesis patterns
• MIDI-driven visual generation
• Spectral analysis visualization
• Beat detection algorithms

g) /wiki/data/experimental/networking.json — multi-system networking:

• TouchDesigner OSC server/client setup
• WebSocket DAT for browser integration
• NDI video streaming patterns
• Shared memory CHOP/TOP for inter-process
• Multi-instance synchronization via TDAbleton

h) /wiki/data/experimental/python-advanced.json — advanced Python in TD:

• Asyncio patterns in TD (run() coroutines)
• Custom extensions (tdu.Dependency)
• Python threading with callbacks
• External library integration (numpy, scipy, opencv)
• Custom DAT operators

Update /wiki/data/experimental/index.json with accurate totalTechniques count after creating all files.