still-manim is a python library for drawing static graphics of conceptual content in domains like math and programming. It's based on 3blue1brown's animation library, manim, except it is designed for static graphics and for running in the browser. The library outputs SVGs.

Try it in the web editor (source code here). Documentation

Example 1:

# this code creates the logo SVG at the top of this README
from smanim import *
stroke_width = 1
lemon = VGroup()
c = Circle(fill_color=YELLOW_D, stroke_color=WHITE, stroke_width=stroke_width)
c.stretch(1.4, dim=0).rotate(PI / 8)
arc = Arc(
    angle=-PI, fill_color=YELLOW_E, stroke_color=WHITE, stroke_width=stroke_width
)
arc.stretch(2, dim=1)
arc.stretch(1.4, dim=0).rotate(PI / 8)
lemon.add(c, arc)
lemon.bring_to_front(c)

spikes = VGroup()
for prop in range(8):
    line = Line(ORIGIN, c.point_from_proportion(prop / 8))
    spikes.add(line)
lemon.add(spikes)

other_lemon: VGroup = lemon.copy()
other_lemon.scale(0.8).shift(RIGHT * 0.5).rotate(-PI / 4)
lemon.shift(LEFT * 0.5)
other_lemon.shift(RIGHT * 0.8)
other_lemon.set_z_index(-10)

lemons = VGroup(lemon, other_lemon)
title = Text("Still Manim", font_size=H1_FONT_SIZE)
lemons.scale_to_fit_height(title.height)
title.next_to(lemons, buff=0.05)
canvas.add(lemons, title)

canvas.draw(crop=True)
# use canvas.snapshot(crop=True) instead if you are running locally

Example 2:

from smanim import *
canvas.set_dimensions(6, 6)
WEIGHTED_GRAPH1 = {
    0: [(1, 2), (2, 1)],
    1: [(2, 5), (3, 11), (4, 3)],
    2: [(5, 15)],
    3: [(4, 2)],
    4: [(2, 1), (5, 4), (6, 5)],
    5: [],
    6: [(3, 1), (5, 1)],
}
vertices, edges, edge_labels = WeightedGraph.from_adjacency_list(WEIGHTED_GRAPH1)
graph = WeightedGraph(
        vertices,
        edges,
        vertex_config={"fill_color": GRAY, "radius": 0.2},
        edge_labels=edge_labels,
        edge_type=Arrow,
        layout_config={"seed": 2},
        include_vertex_labels=True,
    )
start_vertex = graph.vertices[0]
start_vertex.set_color(RED)
pointer = Arrow.points_at(start_vertex, direction=LEFT, color=RED, length=0.5)
start_text = Text("start", color=RED).next_to(pointer)
canvas.add(graph, pointer, start_text)

canvas.draw(crop=True)
# use canvas.snapshot(crop=True) instead  if you are running locally

• Run in the browser (must be pure python or packages easily supported by pyodide, unlike original manim)
• Allow flexible and clear construction of domain-specific graphical objects in as little python code as possible
• Present language constructs to an LLM tutor, understanding that an LLM might not have great spatial awareness
• Provide sensible support for human-AI interaction while editing diagrams (for example, by tracking the line of code for each variable assignment to a mobject and storing that in the mobject element in the SVG.)

• Polygons: polygons, regular polygons, squares, rectangles, triangles
• Arcs: circles, dots
• Lines: line segments, arrows, vectors
• Graphs: directed graphs, undirected graphs, weighted graphs
• Cartesian Graphs: number lines, 2D cartesian graphs, functions for those graphs (Note: The API is very unstable here)
• Text: font family (Roboto or computer-modern), font size, font color, font opacity, font decoration, bold, italics
• Grouping: groups, box lists
• Annotations: crosses, labels, surrounding rectangles

still-manim does not support all the different types of Mobjects in manim. Also, still-manim is entirely in 2D for now.

All constructed mobjects permit:

• spatial relations: e.g., obj1.next_to(obj2))
• spatial transformations: e.g., obj1.rotate(PI / 2)
• styling: e.g., c = Circle(stroke_color=RED)
• grouping: e.g., g = Group(Circle(), Rectangle().shift(RIGHT * 2))
• layering: e.g., circle.set_z_index(10)
• various compositions of mobjects: e.g. a = Arrow.points_at(circle) or circle.add_label(Text("A Circle"))

pip install still-manim

1. Clone this repo:

git clone https://github.com/tommy11jo/still-manim.git

2. Install the deps the first time you open this repo:

poetry install

3. Init the python env each time you open this repo:

poetry shell

5. Run your first still-manim program by navigating to the examples folder and running:

python3 hello_world.py

• Examples can be found in the web editor .
• The API exposed to users for this project is loosely based on the Manim Community's variant of manim.