graphtty 0.0.3

Render directed graphs as ASCII/Unicode art for console output

graphtty

Render directed graphs as ASCII/Unicode art for console output.

Installation

pip install graphtty

Screenshots

React Agent (monokai theme)

Deep Agent (ocean theme)

Workflow Agent (forest theme)

Supervisor Agent (dracula theme)

CLI

Render any graph JSON file straight from the terminal:

graphtty graph.json
graphtty graph.json --theme monokai
graphtty graph.json --ascii
graphtty graph.json --no-types

Options

usage: graphtty [-h] [-t THEME] [--ascii] [--no-types] [--list-themes] [file]

positional arguments:
  file                  Path to a graph JSON file

options:
  -h, --help            show this help message and exit
  -t THEME, --theme THEME
                        Color theme (use --list-themes to see options)
  --ascii               Use plain ASCII instead of Unicode box-drawing characters
  --no-types            Hide node type labels
  --list-themes         List available themes and exit

Themes

Ten built-in color themes that style nodes by type:

graphtty graph.json --theme monokai
graphtty graph.json --theme ocean
graphtty graph.json --theme forest
graphtty graph.json --theme dracula
graphtty graph.json --theme solarized
graphtty graph.json --theme nord
graphtty graph.json --theme catppuccin
graphtty graph.json --theme gruvbox
graphtty graph.json --theme tokyo-night
graphtty graph.json                     # default (no color)

List all available themes:

graphtty --list-themes

Python API

Simple linear chain

from graphtty import AsciiGraph, AsciiNode, AsciiEdge, render

graph = AsciiGraph(
    nodes=[
        AsciiNode(id="a", name="Start", type="entry"),
        AsciiNode(id="b", name="Process", type="tool"),
        AsciiNode(id="c", name="End", type="exit"),
    ],
    edges=[
        AsciiEdge(source="a", target="b"),
        AsciiEdge(source="b", target="c"),
    ],
)

print(render(graph))

  ┌ entry ──┐
  │  Start  │
  └────┬────┘
       │
       ▼
  ┌ tool ───┐
  │ Process │
  └─────┬───┘
        │
        ▼
   ┌ exit ──┐
   │  End   │
   └────────┘

Branching with edge labels

graph = AsciiGraph(
    nodes=[
        AsciiNode(id="start", name="User Input", type="entry"),
        AsciiNode(id="router", name="Router", type="condition"),
        AsciiNode(id="search", name="Web Search", type="tool"),
        AsciiNode(id="calc", name="Calculator", type="tool"),
        AsciiNode(id="respond", name="Respond", type="model"),
    ],
    edges=[
        AsciiEdge(source="start", target="router"),
        AsciiEdge(source="router", target="search", label="search"),
        AsciiEdge(source="router", target="calc", label="calc"),
        AsciiEdge(source="search", target="respond"),
        AsciiEdge(source="calc", target="respond"),
    ],
)

print(render(graph))

           ┌ entry ─────┐
           │ User Input │
           └─────┬──────┘
                 │
                 ▼
          ┌ condition ──┐
          │   Router    │
          └──────┬──────┘
         ┌search─└──calc───┐
         ▼                 ▼
  ┌ tool ──────┐    ┌ tool ──────┐
  │ Web Search │    │ Calculator │
  └──────┬─────┘    └──────┬─────┘
         └───────┌─────────┘
                 ▼
            ┌ model ──┐
            │ Respond │
            └─────────┘

Nested subgraphs

sub = AsciiGraph(
    nodes=[
        AsciiNode(id="s1", name="Fetch Data", type="tool"),
        AsciiNode(id="s2", name="Transform", type="action"),
    ],
    edges=[AsciiEdge(source="s1", target="s2")],
)

graph = AsciiGraph(
    nodes=[
        AsciiNode(id="a", name="Start", type="entry"),
        AsciiNode(id="b", name="ETL Pipeline", type="subgraph", subgraph=sub),
        AsciiNode(id="c", name="Done", type="exit"),
    ],
    edges=[
        AsciiEdge(source="a", target="b"),
        AsciiEdge(source="b", target="c"),
    ],
)

print(render(graph))

     ┌ entry ──┐
     │  Start  │
     └─────┬───┘
           │
           ▼
  ┌ subgraph ──────┐
  │  ETL Pipeline  │
  │ ┌ tool ──────┐ │
  │ │ Fetch Data │ │
  │ └──────┬─────┘ │
  │        │       │
  │        ▼       │
  │  ┌ action ───┐ │
  │  │ Transform │ │
  │  └───────────┘ │
  └────────┬───────┘
           │
           ▼
      ┌ exit ──┐
      │  Done  │
      └────────┘

Node descriptions

graph = AsciiGraph(
    nodes=[
        AsciiNode(id="a", name="LLM", type="model", description="gpt-4o"),
    ],
)

print(render(graph))

Themes (Python API)

from graphtty import render, RenderOptions, get_theme

options = RenderOptions(theme=get_theme("monokai"))
print(render(graph, options))

Render options

from graphtty import RenderOptions

options = RenderOptions(
    use_unicode=True,   # Use Unicode box-drawing characters (default: True)
    show_types=True,    # Show node types in box borders (default: True)
    padding=2,          # Padding around the diagram (default: 2)
)

print(render(graph, options))

JSON format

graphtty reads a simple JSON format:

{
  "nodes": [
    { "id": "a", "name": "Start", "type": "entry" },
    { "id": "b", "name": "Process", "type": "tool", "description": "optional detail" },
    { "id": "c", "name": "End", "type": "exit" }
  ],
  "edges": [
    { "source": "a", "target": "b", "label": "optional" },
    { "source": "b", "target": "c" }
  ]
}

Testing

uv run pytest tests/ -v

Acknowledgements

Built with :heart: on top of grandalf, a Python library for graph layout using the Sugiyama algorithm.