📝 AI Article Generator

Project Overview

AI Article Generator is a simple, modular content generation web app built with LangChain.js and LangChain Expression Language (LCEL) . It allows users to input a topic and receive a fully structured article, including:

• Headline
• Introduction
• Body
• Conclusion

The app combines a React frontend and a Node.js + LangChain backend using LCEL's .pipe() syntax to create a smooth content generation experience.

---

How It Works

Flow Summary

1. User Input : Enter a topic in the web UI.
2. Processing (Backend) :

• LCEL pipelines generate each article section in order.

1. Output : Display the complete article in the browser.

---

Modifications by wildfirebill

The following changes were made to the original project:

Backend (backend/src/index.ts)

• Replaced LangChain ChatOpenAI (Groq) with direct Ollama HTTP API calls
• Added GET /models endpoint to list available Ollama models
• POST /generate-article now accepts model and baseUrl from the request body
• Added SEO tags generation as a fifth article section
• Removed dotenv, @langchain/core, @langchain/openai dependencies

Frontend (frontend/src/App.tsx)

• Added collapsible Settings panel with Ollama URL input and model picker
• Settings persist across page reloads via localStorage
• Added light/dark mode toggle with CSS custom properties
• Dashboard displays SEO tags as pill badges below the article
• Added error display for failed requests

Frontend (frontend/src/index.css, frontend/src/App.css)

• Rewrote styles to use CSS custom properties for theme support
• Added dark theme variables under [data-theme="dark"]
• Added settings panel, tag pills, and theme toggle styles

Root

• Added README.md with setup instructions and modification log