Gradio: Create Interactive ML Demos That Anyone Can Use in 5 Lines

Updated
7 min read
Gradio: Create Interactive ML Demos That Anyone Can Use in 5 Lines
TL;DR

Gradio lets you wrap any Python function in a web UI with automatic type inference, and every app doubles as a REST API - deploy to Hugging Face Spaces for free in minutes.

What Gradio Is and Why It Matters

Gradio is the fastest way to create interactive demos for ML models. Five lines of Python creates a shareable web UI with inputs, outputs, and an automatically generated REST API.

The library is maintained by Hugging Face and is the standard for sharing model demos in the ML community. Every model on Hugging Face Spaces is a Gradio (or Streamlit) app.

gr.Interface: Simplest Path to a Demo

python
import gradio as gr
from transformers import pipeline

classifier = pipeline("sentiment-analysis")

def analyze(text: str) -> dict:
    result = classifier(text)[0]
    return {"label": result["label"], "confidence": round(result["score"], 3)}

demo = gr.Interface(
    fn=analyze,
    inputs=gr.Textbox(label="Input text", placeholder="Type something..."),
    outputs=gr.JSON(label="Result"),
    title="Sentiment Analyzer",
    examples=["I love this product!", "This is terrible."],
)

demo.launch()

Gradio infers input/output types automatically. Pass a string → gets a Textbox. Pass an image → gets an Image upload component.

gr.Blocks: Custom Layouts

python
import gradio as gr

with gr.Blocks() as demo:
    gr.Markdown("# Image Classifier")

    with gr.Row():
        image_input = gr.Image(type="pil", label="Upload image")
        output_label = gr.Label(num_top_classes=5, label="Predictions")

    with gr.Row():
        classify_btn = gr.Button("Classify", variant="primary")
        clear_btn = gr.ClearButton([image_input, output_label])

    classify_btn.click(fn=classify_image, inputs=image_input, outputs=output_label)

demo.launch()

gr.ChatInterface for LLM Chatbots

python
import gradio as gr
from anthropic import Anthropic

client = Anthropic()

def chat(message: str, history: list) -> str:
    messages = [{"role": h["role"], "content": h["content"]} for h in history]
    messages.append({"role": "user", "content": message})

    response = client.messages.create(
        model="claude-3-haiku-20240307",
        max_tokens=1024,
        messages=messages,
    )
    return response.content[0].text

demo = gr.ChatInterface(fn=chat, title="Claude Chat", type="messages")
demo.launch()

For streaming, return a generator from your chat function and Gradio handles the streaming UI automatically.

Auto-Generated API

Every Gradio app exposes a REST API automatically:

bash
# Call your Gradio app's API
curl -X POST http://localhost:7860/api/predict   -H "Content-Type: application/json"   -d '{"data": ["Hello world"]}'

The Python client makes this even simpler:

python
from gradio_client import Client

client = Client("https://your-app.hf.space")
result = client.predict("Hello world", api_name="/predict")

Sharing Your Demo

python
# Temporary public URL (72 hours)
demo.launch(share=True)

# Deploy to Hugging Face Spaces (permanent, free)
# 1. Create a Space at huggingface.co/spaces
# 2. Push your app.py + requirements.txt to the Space repo
# 3. HF Spaces builds and hosts it automatically

Gradio vs Streamlit

Use Gradio when: you want to demo a specific ML model, you want an auto-generated API, you are deploying to Hugging Face Spaces, or you need the simplest possible UI. Use Streamlit when: you are building a full data app (dashboards, multi-page), you need custom layouts and complex state, or you are building an internal tool.

Resources: Gradio docs, GitHub, HF Spaces.

#gradio#demo#huggingface-spaces#ui#python

One AI engineering post, weekly

LLM benchmarks, prompt techniques, and token-cost breakdowns, delivered to your inbox.

What should we test or explain next?

Comments

Have a question or something to add? Sign in to join the discussion.

Mahmudul Haque Qudrati — CEO & ML Engineer at Pristren

Mahmudul Haque Qudrati

CEO & ML Engineer

Software architect, full-stack engineer, and machine learning practitioner. Founder & CEO of Pristren, CTO at Venture Kick-backed PIN Sport Application GmbH (Basel, Switzerland), Software Development Lead at DailyFrame GmbH (building Dimetra patient transport software), and creator of Zlyqor. Focused on custom workflow automation, scalable web and mobile applications, and production AI/ML systems.