alivirgo/Major-AI-Skills / plugins/mas-3d-games/skills/cinema-4d
cinema-4d
VerifiedScript Cinema 4D scenes with c4dpy and the Maxon API, wire MoGraph/Fields systems, and debug Redshift GPU renders, AOVs, and headless Commandline farm jobs agents routinely miss.
Maxon Cinema 4D AI Skill Guide (Claude)
Overview & Engine Architecture
Maxon Cinema 4D (C4D 2024/2025+) is a motion-design and VFX DCC. Agents should treat scenes as versioned project packages (assets + caches + render settings), MoGraph as procedural contracts (cloners + effectors + fields), and Redshift as a GPU memory budget problem—not “turn samples up.” Automation runs through c4dpy, the classic c4d API, the maxon graph API (Scene Nodes / Neutron), and headless Commandline renders.
┌─────────────────────────────────────────────────────────────┐
│ C4D build → cache → render loop │
│ │
│ Authoring │
│ ├── Classic OM + Scene Nodes / Neutron │
│ ├── MoGraph: Cloner → Effectors → Fields │
│ └── Unified Sim: Pyro / Cloth / Soft / Rigid │
│ │
│ Automation │
│ ├── c4dpy (headless Python 3) │
│ ├── c4d + maxon APIs │
│ └── Commandline.exe farm flags (-render -frame -oimage) │
│ │
│ Redshift │
│ ├── GPU VRAM + Out-of-Core + texture streaming │
│ └── ACEScg OCIO, AOVs / Cryptomatte, .rs proxies │
└─────────────────────────────────────────────────────────────┘
Version pins: document Cinema 4D year + build and Redshift build in every pipeline README. Plugin IDs and Python APIs drift across major releases—never assume R20 snippets still compile.
When to use / when not to
Use when
- Authoring or debugging MoGraph systems, Takes, and Redshift materials/AOVs.
- Headless batch export (USD/Alembic/FBX) or farm renders via Commandline / c4dpy.
- Packaging projects, fixing missing textures, or remapping asset paths for a render node.
Do not use when
- Heavy procedural FX / USD LOPs / Karma is the center of the pipeline →
@houdini. - Organic high-poly sculpt primary →
@zbrushthen bring mesh into C4D. - Real-time game runtime packaging → Unreal/Unity skills after export.
Operational Capabilities & Agent Directives
- Pin C4D + Redshift builds in CI/farm docs; log
c4dpy/ Commandline version at job start. - Prefer data API (
documents,BaseObject, tags) over GUI-only ops in headless scripts; callc4d.EventAdd()only when a GUI session needs refresh. - Always
KillDocument(doc)after headless work to avoid batch memory leaks. - Redshift VRAM first: proxies (
.rs), texture streaming / TX, Out-of-Core—before raising samples. - Color: ACEScg / OCIO must match farm workers; linear EXR for comp, display transforms only for preview.
- MoGraph dynamics: fix collision shape + steps-per-frame before blaming effectors.
- Exit non-zero on failed LoadDocument / SaveDocument / render so farm schedulers fail the task.
- Never invent license servers, registry keys, or credentials—ask for authorized farm paths and existing env.
Production Example: headless USD export + Redshift still gate
batch_export_usd.py — run via c4dpy:
"""Cinema 4D headless USD export.
Usage: c4dpy batch_export_usd.py <scene.c4d> <output_dir>
"""
from __future__ import annotations
import os
import sys
import c4d
from c4d import documents
USD_EXPORT_ID = 1040540 # verify against installed C4D build
def export_usd(scene_path: str, output_dir: str) -> None:
if not os.path.isfile(scene_path):
print(f"C4D_PIPELINE_FAIL missing scene: {scene_path}", file=sys.stderr)
sys.exit(1)
doc = documents.LoadDocument(
scene_path,
c4d.SCENEFILTER_OBJECTS | c4d.SCENEFILTER_MATERIALS,
)
if doc is None:
print(f"C4D_PIPELINE_FAIL LoadDocument: {scene_path}", file=sys.stderr)
sys.exit(1)
documents.SetActiveDocument(doc)
os.makedirs(output_dir, exist_ok=True)
stem = os.path.splitext(os.path.basename(scene_path))[0]
usd_path = os.path.join(output_dir, f"{stem}.usd")
ok = documents.SaveDocument(
doc,
usd_path,
c4d.SAVEDOCUMENTFLAGS_DONTADDTORECENTLIST,
USD_EXPORT_ID,
)
documents.KillDocument(doc)
if not ok:
print(f"C4D_PIPELINE_FAIL USD export: {usd_path}", file=sys.stderr)
sys.exit(1)
print(f"exported {usd_path}")
if __name__ == "__main__":
if len(sys.argv) < 3:
print("Usage: c4dpy batch_export_usd.py <scene.c4d> <output_dir>", file=sys.stderr)
sys.exit(2)
export_usd(sys.argv[1], sys.argv[2])
Validation / farm render loop:
# Version gate
"C:\Program Files\Maxon Cinema 4D 2025\c4dpy.exe" -c "import c4d; print(c4d.GetC4DVersion())"
# Asset export
"C:\Program Files\Maxon Cinema 4D 2025\c4dpy.exe" pipeline\batch_export_usd.py scene.c4d out\
# Headless Redshift stills (paths are install-local — do not invent farm credentials)
"C:\Program Files\Maxon Cinema 4D 2025\Commandline.exe" -render "D:\Jobs\scene.c4d" -frame 1 1 -oimage "D:\Renders\beauty_####"
Technical Troubleshooting Matrix
| Signature | Likely cause | Fix path |
|---|---|---|
CUDA_ERROR_OUT_OF_MEMORY / Redshift abort | Textures + tessellation exceed VRAM | Out-of-Core + texture streaming; convert heavy meshes to .rs proxies; lower subdivision |
| Alembic/FBX materials missing | Face sets / selection names do not match | Match Polygon Selection Tags to Alembic part names; reassign RS materials to tags |
| MoGraph dynamics explode frame 0 | Overlapping cloners / bad collision shape | Multi-Box or Convex Hull; raise collision margin + Project Dynamics steps |
AttributeError on c4d.* | Deprecated R20 API vs 2024/2025 | Prefer current docs; use maxon for graph/node work |
| Farm black frames | Missing camera, lights, or texture paths | Assert active camera; Project Asset Inspector; relative paths + collect assets |
| OCIO / ACES mismatch local vs farm | Different OCIO config | Pin OCIO path in env on all workers; compare EXR linear vs preview LUT |
| Commandline ignores frame/output | Wrong flag order / wrong binary | Use Commandline.exe (not GUI); verify -render -frame -oimage |
Best Practices
- Collect assets before farm submit; prefer relative project paths.
- Cache MoGraph / sims to disk before long renders; do not recook on every frame if unchanged.
- Name Takes and cameras explicitly; pass
-camera/ take flags on the farm. - Export multi-pass EXR AOVs + Cryptomatte for comp—not flattened preview PNG as hero deliverable.
- Keep Python scripts idempotent: find objects by name before creating duplicates.
- Document required env (Redshift cache dir, OCIO) in the shot README.
Limitations
- Scene Nodes / Neutron Python surfaces differ from classic OM—verify against the installed build’s Maxon API.
- Redshift feature parity and VRAM behavior depend on GPU vendor/driver; CPU fallback is not a free farm substitute.
- Plugin IDs (USD export, etc.) can change—resolve via
plugins.FindPluginwhen uncertain. - Stop and ask if license/farm credentials, proprietary plugin paths, or client color policies are missing.
Related Skills
@houdini— heavy FX, Solaris/Karma, PDG wedging@blender— open-source bpy / Cycles automation@zbrush— high-res organic sculpt → retopo into C4D@substance-painter— PBR texturing after mesh export@unreal-engine/@unity— real-time destinations after USD/FBX
Agent Operational Directive
MANDATORY: Pin C4D and Redshift builds. In headless scripts always
KillDocumentand exit non-zero on failure. Fix Redshift VRAM (proxies, streaming, Out-of-Core) before cranking samples. Match OCIO/ACES on every farm worker. Never invent licenses, registry secrets, or credentialed paths.
Source anchors (research)
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| finishing-a-development-branchobra/superpowersUse when implementation is complete, all tests pass, and you need to decide how to integrate the work | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| systematic-debuggingobra/superpowersUse when encountering any bug, test failure, or unexpected behavior, before proposing fixes | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| test-driven-developmentobra/superpowersUse when implementing any feature or bugfix, before writing implementation code | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| skill-creatoranthropics/skillsCreate new skills, modify and improve existing skills, and measure skill performance. Use when users want to create a skill from scratch, edit, or optimize an existing skill, run evals to test a skill, benchmark skill performance with variance analysis, or optimize a skill's description for better triggering accuracy. | Stars: 198k | Updated | Claude CodeCodex | Verified | Install |