연결

After Effects Vision

VolksRat71byVolksRat71
•
9
•
개발자 도구
생산성 및 워크플로
디자인 도구

Drive live Adobe After Effects and Illustrator sessions with agent-driven editing and visual pixel feedback for iterative design refinement.

After Effects Vision bridges multimodal AI models and Adobe Creative Cloud via a native CEP extension and MCP server, allowing agents to manipulate live compositions and artboards and visually inspect what they create. By capturing rendered frames, contact sheets, or isolated layers directly into the agent's context, models can verify visual output and autonomously correct errors. With 16 tools for After Effects and 10 for Illustrator, all operations reference stable IDs and match names across undo-safe batches, supporting workflows from automated typography and layout to complex rotoscoping round-trips with external AI vision models.

주요 기능

019 GitHub stars
02Visual feedback loop enabling agents to capture frames, contact sheets, or isolated layers to verify their output
0316 After Effects tools covering layer manipulation, keyframing, shape creation, masking, and rendering
0410 Illustrator tools for vector geometry, artboards, typography, and automated alignment
05Stable element addressing using internal IDs and matchName paths to survive reordering and localized UI versions
06Atomic undo grouping ensuring multi-step agent batches can be reverted in a single undo step

사용 사례

01AI-assisted rotoscoping round trips that import footage, run object tracking models, and write back animated masks
02Autonomous motion graphics creation and animation with continuous visual verification in After Effects
03Programmatic vector illustration, artboard layout, and asset export generation inside Adobe Illustrator