manim-skill

majiayu000's avatarfrom majiayu000

Render Manim scenes to still frames and review them to diagnose animation/layout issues. Use when debugging scenes, verifying visual correctness, or inspecting frame-by-frame output.

⭐0stars🔀0forks📁View on GitHub🕐Updated Jan 5, 2026

When & Why to Use This Skill

This Claude skill streamlines the creation and debugging of Manim (Mathematical Animation) scenes by enabling frame-by-frame rendering and visual inspection. It allows developers and creators to diagnose layout issues, verify timing, and inspect transitions using contact sheets and high-resolution stills, significantly reducing the feedback loop compared to full video rendering.

Use Cases

  • Debugging Animation Layouts: Quickly render specific frames to identify and fix overlapping objects or incorrect positioning during complex scene transitions.
  • Visual Quality Assurance: Use generated contact sheets (4x4 grids) to get a comprehensive overview of animation progression and ensure visual consistency.
  • High-Resolution Inspection: Review individual PNG frames at full resolution to catch subtle rendering glitches or cropping issues that might be missed in low-quality video previews.
  • Iterative Scene Development: Rapidly test code changes by refreshing frames at variable FPS, allowing for fast-paced debugging of mathematical visualizations.
namemanim-skill
descriptionRender Manim scenes to still frames and review them to diagnose animation/layout issues. Use when debugging scenes, verifying visual correctness, or inspecting frame-by-frame output.

Manim Frame Review

Render a Manim scene to PNG frames, inspect them visually, and iterate quickly.

Workflow

1. Render frames

Use the helper script to clear stale frames and re-render in one step:

python3 scripts/refresh_frames.py [--fps FPS] [--contact-sheet] <file.py> <SceneName>
  • --fps FPS: Frames per second (default: 2). Increase for fast-paced animations.
  • --contact-sheet: Generate a contact sheet (4x4 grid with evenly sampled frames) after rendering.

Or render manually:

manim -ql --fps 2 --format=png --silent <file>.py <SceneName> | tail -n 5

Frames are written to media/images/<file>/.

2. Inspect frames

For a quick overview, use --contact-sheet when rendering, producing /media/images/<file>/<SceneName>_sheet.png. Always read the contact sheet first to get an overview of the animation progression. Then read at least one full-size frame (e.g., the first, middle, or a frame showing a key transition) to verify details at full resolution. To regenerate a contact sheet without re-rendering:

python3 scripts/make_contact_sheet.py <frames_directory> --scene-name <SceneName>

Alternatively, read frame images directly for visual inspection. Check early, middle, and late frames plus key transitions to catch overlaps, cropping, or timing issues.

For fast-paced and detailed animations, it may be necessary to increase --fps to capture more frames, for example --fps 4 or --fps 8. For example, if you were previously inspecting frames 0005 and 0006 at fps 2 and need to know what happens in-between, at fps 4 you would inspect frames 0010 and 0012 instead and have a new frame 0011 in between.

3. Iterate

Fix issues in the scene code, then repeat steps 1-2 until correct.

4. Render video

When the user wants the full video:

manim -ql <file>.py <SceneName>

Key render options:

  • -q [l|m|h|p|k]: Quality (l: 854x480@15fps, m: 1280x720@30fps, h: 1920x1080@60fps)
  • --format [png|gif|mp4|webm|mov]: Output format (default: mp4)
  • -r W,H: Custom resolution
  • --enable_gui: Enable GUI interaction (use when explicitly requested)

References

Search references/manim-docs/ recursively for Manim documentation (tutorials, guides, reference, FAQ).