Neura Market Manim Animation Reference

Overview

This reference documents the complete pipeline for producing 3Blue1Brown-style animated explainer videos using Manim Community Edition. The tool is designed for educational, mathematical, and technical content, including animated explanations, math animations, concept visualizations, algorithm walkthroughs, and technical explainers. It handles programmatic animation with geometric or mathematical content, algorithm visualizations, equation derivations, architecture diagrams, data stories, paper explainers, and 3D visualizations.

Prerequisites

Before beginning any project, ensure the following dependencies are installed:

• Python 3.10 or later
• Manim Community Edition v0.20 or later (install with pip install manim)
• A LaTeX distribution: texlive-full on Linux, mactex on macOS
• ffmpeg

Run scripts/setup.sh to verify all dependencies are present.

Project Structure

Every project must follow this directory layout:

project-name/
  plan.md                # Narrative arc, scene breakdown
  script.py              # All scenes in one file
  concat.txt             # ffmpeg scene list
  final.mp4              # Stitched output
  media/                 # Auto-generated by Manim
    videos/script/480p15/

Full Video Production Pipeline

The pipeline follows six ordered steps:

PLAN --> CODE --> RENDER --> STITCH --> AUDIO (optional) --> REVIEW

Step 1: Plan

Write plan.md containing the narrative arc, scene list, visual elements, color palette, and voiceover script. Before coding, articulate the narrative arc: what misconception is corrected, what is the "aha moment", and what visual story takes the viewer from confusion to understanding.

Step 2: Code

Write script.py with one class per scene, each independently renderable. Use shared color constants at the file top. Set self.camera.background_color in every scene. Add subcaptions on every animation. End each scene with FadeOut of all mobjects.

Use this template for every scene file:

from manim import *

BG = "#1C1C1C"
PRIMARY = "#58C4DD"
SECONDARY = "#83C167"
ACCENT = "#FFFF00"
MONO = "Menlo"

class Scene1_Introduction(Scene):
    def construct(self):
        self.camera.background_color = BG
        title = Text("Why Does This Work?", font_size=48, color=PRIMARY, weight=BOLD, font=MONO)
        self.add_subcaption("Why does this work?", duration=2)
        self.play(Write(title), run_time=1.5)
        self.wait(1.0)
        self.play(FadeOut(title), run_time=0.5)

Step 3: Render

Run Manim with the appropriate quality flag. Always iterate at -ql quality for drafts; only use -qh for final output.

manim -ql script.py Scene1_Introduction Scene2_CoreConcept  # draft
manim -qh script.py Scene1_Introduction Scene2_CoreConcept  # production

Step 4: Stitch

Create concat.txt listing scene MP4 files in order, then run ffmpeg concat to produce final.mp4.

cat > concat.txt << 'EOF'
file 'media/videos/script/480p15/Scene1_Introduction.mp4'
file 'media/videos/script/480p15/Scene2_CoreConcept.mp4'
EOF
ffmpeg -y -f concat -safe 0 -i concat.txt -c copy final.mp4

Step 5: Audio (Optional)

Add voiceover and/or background music via ffmpeg. Voiceover can be generated using ElevenLabs or Qwen3-TTS.

Step 6: Review

Render preview stills without full video render to verify against the plan.

manim -ql --format=png -s script.py Scene2_CoreConcept  # preview still

Parameters

Quality Flag

The quality flag determines resolution and FPS:

• -ql: 854x480 at 15 FPS (draft)
• -qm: 1280x720 at 30 FPS (medium)
• -qh: 1920x1080 at 60 FPS (production)

run_time (Animation Speed)

Duration of an animation in seconds. Standard values:

• Title/intro: 1.5s
• Key equation reveal: 2.0s
• Transform/morph: 1.5s
• Supporting label: 0.8s
• FadeOut cleanup: 0.5s
• "Aha moment" reveal: 2.5s

self.wait() Duration

Pause after an animation. Standard values:

• After title/intro: 1.0s
• After key equation reveal: 2.0s
• After transform/morph: 1.5s
• After supporting label: 0.5s
• After FadeOut cleanup: 0.3s
• After "aha moment" reveal: 3.0s

font_size

Text size in Manim units. Scale:

• Title: 48
• Heading: 36
• Body: 30
• Label: 24
• Caption: 20

Minimum is font_size=18 for readability.

buff (For Edge Text)

Padding when placing text at screen edge. Must be 0.5 or greater.

label.to_edge(DOWN, buff=0.5)  # never < 0.5

subcaption

Text for voiceover or subtitle synced to an animation. Can be passed as argument to self.play() or via self.add_subcaption().

self.add_subcaption("text", duration=N)
# or
self.play(..., subcaption="text")

Typography and Colors

Font

Use monospace fonts for all text. Manim's Pango renderer produces broken kerning with proportional fonts at all sizes. Define the font once at the top of the file:

MONO = "Menlo"  # define once at top of file

Text("Fourier Series", font_size=48, font=MONO, weight=BOLD)  # titles
Text("n=1: sin(x)", font_size=20, font=MONO)                  # labels
MathTex(r"\nabla L")                                            # math (uses LaTeX)

Color Palette

All scenes in a project must share a cohesive color palette. Standard palettes include:

• Dark theme: #1C1C1C background, #58C4DD primary, #83C167 secondary, #FFFF00 accent
• Alternative dark: #2D2B55 background, #FF6B6B primary, #FFD93D secondary, #6BCB77 accent
• Neon: #0A0A0A background, #00F5FF primary, #FF00FF secondary, #39FF14 accent
• Light: #1A1A2E background, #EAEAEA primary, #888888 secondary, #FFFFFF accent

For visual design guidance, see references/visual-design.md.

Per-Scene Variation Requirement

Every scene must vary from the others in these four ways:

1. Choose a different dominant color from the palette for each scene.
2. Use a different layout (not always center everything).
3. Use a different animation entry method (Write, FadeIn, GrowFromCenter, Create).
4. Vary visual weight: some scenes dense, others sparse.

Creative Divergence (Experimental/Creative Requests Only)

For experimental or creative requests, select one of two strategies:

SCAMPER

Apply one of these operations to the standard visualization metaphor: Substitute, Combine, Reverse, Modify, or Eliminate.

Assumption Reversal

List standard assumptions about how the topic is visualized, pick the most fundamental one, reverse it, and explore what the reversal reveals.

LaTeX Handling

Always use raw strings for LaTeX. Never use regular strings.

# WRONG: MathTex("\frac{1}{2}")
# RIGHT:
MathTex(r"\frac{1}{2}")

Animation Rules

Mobject Management

Never animate a mobject that hasn't been added to the scene.

self.play(Create(circle))  # must add first
self.play(circle.animate.set_color(RED))  # then animate

Text Replacement

FadeOut old text before writing new text in the same position. Use ReplacementTransform instead of Write on top.

self.play(ReplacementTransform(note1, note2))  # not Write(note2) on top

Opacity Layering

• Primary elements: 1.0
• Contextual elements: 0.4
• Structural elements (axes, grids): 0.15

Breathing Room

Every animation must be followed by self.wait() for breathing room. A 2-second pause after a key reveal is never wasted.

Scene Cleanup

End each scene with a fade out of all mobjects:

self.play(FadeOut(Group(*self.mobjects)))

Content Principles

• Show geometry before algebra: show the shape first, the equation second.
• Use .animate for property transitions on existing mobjects (see references/animations.md).
• For algorithm visualizations, see references/graphs-and-data.md.
• For equation derivations, see references/equations.md.
• For architecture diagrams, see references/mobjects.md.
• For 3D visualizations, see references/camera-and-3d.md.
• For paper explainers, see references/paper-explainer.md.
• For scene planning, see references/scene-planning.md.
• For decorations, see references/decorations.md.
• For updaters and trackers, see references/updaters-and-trackers.md.
• For production quality, see references/production-quality.md.
• For animation design thinking, see references/animation-design-thinking.md.
• For troubleshooting, see references/troubleshooting.md.
• For rendering details, see references/rendering.md.

Failure Modes

Avoid these common failures:

• Cluttered, poorly timed, or "AI-generated slides" appearance (violates first-render excellence).
• Random different colors across scenes (violates cohesive visual language).
• Broken kerning in text (using proportional fonts instead of monospace).
• LaTeX errors (missing raw string prefix).
• Animation on non-added mobject (causes runtime error).
• Text overlapping or clipped at edge (buff less than 0.5).
• Rushing from one animation to the next without self.wait().
• Missing subcaptions on animations.
• Using Write on top of existing text instead of ReplacementTransform.

Example Modes

The tool supports these modes, each with specific reference documents:

• Concept explainer: topic/concept input, animated explanation with geometric intuition output. See references/scene-planning.md.
• Equation derivation: math expressions input, step-by-step animated proof output. See references/equations.md.
• Algorithm visualization: algorithm description input, step-by-step execution with data structures output. See references/graphs-and-data.md.
• Data story: data/metrics input, animated charts, comparisons, counters output. See references/graphs-and-data.md.
• Architecture diagram: system description input, components building up with connections output. See references/mobjects.md.
• Paper explainer: research paper input, key findings and methods animated output. See references/scene-planning.md.
• 3D visualization: 3D concept input, rotating surfaces, parametric curves, spatial geometry output. See references/camera-and-3d.md.