Skip to content
Chethan Yadav
Work

Engineering/2024/Active

AI Text Summarizer

A simple AI text summarizer app that uses the Hugging Face Inference API to summarize text.

JavaScriptExpressNode.js
Repository

Overview

Built while working through Postman Academy's "AI Text Summarizer App" project, this is a small Express server that wraps the Hugging Face Inference API behind two endpoints — text summarization and text-to-image generation. It exists mainly to keep an API key off the client: the browser never talks to Hugging Face directly, it hits a local Express route, which attaches the token server-side and forwards the request.

The summarizer runs on Facebook's bart-large-cnn model; a second, less-advertised route runs Black Forest Labs' FLUX.1-dev model to turn a text prompt into an image. Both are straightforward POST-in, response-out proxies with no queueing, caching, or rate limiting — appropriate for a learning project, not a production service.

Key Features

  • POST /summarize — sends text_to_summarize to facebook/bart-large-cnn via the Hugging Face Inference API and returns the generated summary
  • POST /texttoimage — sends text_to_imagegen to black-forest-labs/FLUX.1-dev and returns the generated image
  • Static frontend served directly from Express (public/index.html for summarization, public/texttoimage.html for image generation) — no separate frontend build step
  • Minimal dependency footprint: express, axios, dotenv

Installation

git clone https://github.com/chethanyadav456/AI-Text-Summarizer-App.git
cd AI-Text-Summarizer-App
npm install

Copy the example environment file and add a Hugging Face access token:

cp example.env .env
ACCESS_TOKEN=your_hugging_face_access_token

Then start the server:

npm run start

The app listens on http://localhost:3000 and serves the static frontend from public/.

Usage

// Summarize text
fetch('http://localhost:3000/summarize', {
  method: 'POST',
  headers: { 'Content-Type': 'application/json' },
  body: JSON.stringify({ text_to_summarize: 'Long article text goes here...' }),
});
 
// Generate an image from a text prompt
fetch('http://localhost:3000/texttoimage', {
  method: 'POST',
  headers: { 'Content-Type': 'application/json' },
  body: JSON.stringify({ text_to_imagegen: 'A watercolor painting of a lighthouse' }),
});

View on GitHub ↗