Dear PyGui

What Is Dear PyGui?

Dear PyGui is a fast, GPU-accelerated Python GUI framework built on Dear ImGui — the immediate-mode UI library popular in game development and real-time tooling. Instead of retaining widget objects that persist across frames, immediate-mode GUIs rebuild the interface every frame. This sounds wasteful but is extremely efficient for dynamic, data-heavy interfaces that change constantly.

Dear PyGui excels where traditional retained-mode toolkits (tkinter, Qt) struggle: live plots updating at 60 FPS, node editors, debug overlays, and instrument dashboards.

Architecture & Core Concepts

• Immediate mode — UI code runs every frame inside the render loop.
• Context — dpg.create_context() initializes; dpg.destroy_context() cleans up.
• Viewport — the OS window via create_viewport() + show_viewport().
• Windows & items — dpg.window(), add_button(), add_plot() build the UI tree.
• Callbacks — callback parameter on widgets; dpg.set_frame_callback() for per-frame logic.
• GPU rendering — all drawing goes through OpenGL/DirectX for high throughput.

Built-in capabilities

Installation

pip install dearpygui

# Requires Python 3.8+ and a GPU with OpenGL support
python -c "import dearpygui.dearpygui as dpg; print('OK')"

Example 1 — Basic window

import dearpygui.dearpygui as dpg

dpg.create_context()
with dpg.window(label="Demo"):
    dpg.add_text("Hello from Dear PyGui")
    dpg.add_button(label="Click", callback=lambda: print("hi"))
dpg.create_viewport(title="Dear PyGui demo", width=400, height=200)
dpg.setup_dearpygui()
dpg.show_viewport()
dpg.start_dearpygui()
dpg.destroy_context()

Example 2 — Live updating plot

import dearpygui.dearpygui as dpg
import math

dpg.create_context()
with dpg.window(label="Live sine wave"):
    with dpg.plot(label="Plot", height=300, width=500):
        dpg.add_plot_axis(dpg.mvXAxis, label="x")
        y_axis = dpg.add_plot_axis(dpg.mvYAxis, label="y")
        series = dpg.add_line_series([], [], parent=y_axis)

phase = [0.0]
def update():
    phase[0] += 0.05
    xs = [x/10 for x in range(628)]
    ys = [math.sin(x/50 + phase[0]) for x in range(628)]
    dpg.set_value(series, [xs, ys])

dpg.set_frame_callback(1, lambda: dpg.set_frame_callback(dpg.get_frame_count()+1, update) or update())
dpg.create_viewport(title="Plot", width=550, height=400)
dpg.setup_dearpygui(); dpg.show_viewport(); dpg.start_dearpygui(); dpg.destroy_context()

Practical advice

• Think in frames: UI code that runs every frame should stay lightweight.
• Use dpg.add_table() for large data grids with sorting and scrolling.
• Dear PyGui is not ideal for traditional form-based business apps — use Qt or tkinter instead.
• Combine with NumPy for high-performance plot data without Python loops where possible.
• The aesthetic is "developer tool" — embrace it or choose CustomTkinter for polish.

Real-world use cases

• Real-time sensor dashboards and oscilloscope-style displays
• ML training monitors (loss curves, metrics live)
• Game engine debug panels and profilers
• Scientific instrument control interfaces
• Node-based visual programming editors