Lysara

• 

  ✨ Lysara — Neon Purple AI Assistant

✨ Lysara — Neon Purple AI Assistant

Inspiration

Lysara was born from a simple dream: creating something more than just another browser extension. I wanted an AI companion that feels alive, one that can analyze, rewrite, translate, summarize, and proofread, not just copy and paste.

I’ve always been captivated by glowing, futuristic designs, especially neon purple. That aesthetic inspired the look and feel of Lysara. It’s not just functional, it’s artistic, an app that feels immersive and personal every time you use it.

What it does

Lysara is an AI companion that makes working with text effortless and fun. It can analyze any webpage, extract and clean the content, detect the language, count words and characters, and even estimate reading time. Beyond that, it empowers you to transform text in multiple ways: translate it into more than ten languages, summarize long articles into digestible insights, proofread for grammar and style, rewrite with a custom prompt, or generate brand new content from scratch. Everything is wrapped in a glowing neon purple interface that feels alive and inspiring.

How we built it

The web app was built in Python using Streamlit for the interface, combined with libraries like trafilatura for text extraction, deep-translator for multilingual support, transformers and torch for summarization and generation, and language_tool_python for proofreading. I also designed a Chrome extension that connects to the same APIs so users can launch Lysara directly from their browser. For packaging, I set up virtual environments, managed dependencies with requirements.txt, and used PyInstaller to turn the app into a standalone .exe.

Lysara is an AI-powered app with two main worlds. In the first, you can analyze any page, pull out its content, detect the language, count the words and characters, and even get an estimate of how long it would take to read. In the second, you can take that text, or write your own, and transform it. You can translate it into over ten languages, summarize long passages into clear insights, proofread and polish your writing, rewrite with your own style in mind, or let the app generate something brand new for you. All of this lives inside a glowing neon purple interface that makes the whole experience feel sleek and unique.

Challenges we ran into

Getting all the dependencies to play nicely together was a huge challenge. Libraries like lxml, regex, and language_tool_python caused issues across different environments like Google Colab, Jupyter, and VS Code. The proofreader API was particularly tricky since it required upgrades, so I had to implement a fallback correction system to keep things reliable. Packaging the app into a Windows executable took trial and error with PyInstaller, and performance tuning was necessary to keep transformer models from slowing everything down.

Accomplishments that we're proud of

I’m proud that Lysara works smoothly both as a web app and as a Chrome extension. The neon purple interface is unique and sets the app apart, blending utility with artistry. I’m also proud of solving the proofreader fallback problem, it means the app never crashes even if APIs misbehave. Finally, packaging Lysara into a standalone app was a big win because it makes installation and use simple for anyone.

What we learned

This project taught me how to combine multiple NLP tools into one workflow, how to handle failing APIs gracefully, and how to design a consistent user experience across platforms. I deepened my knowledge of Streamlit and Gradio, learned a lot about Python environments, and discovered the nuances of packaging apps into executables. I also got more comfortable debugging dependency issues and managing performance for transformer-based models.

What's next for Lysara

The next step is to expand Lysara with more creative AI tools; for example, style transfer for writing, smarter summarization options, and direct integration with platforms like Google Docs and Notion. I also plan to publish the Chrome extension to the Chrome Web Store, polish the desktop app installer for Windows, and eventually make versions for macOS and Linux. Long term, I want Lysara to feel like a personal AI companion that lives on any device, ready to help you think, create, and express yourself in style.

Built With

• accelerate
• beautifulsoup4
• deep-translator
• google-colab
• gradio
• jupyter-notebook
• langdetect
• language-tool-python
• pyinstaller
• python
• requests
• streamlit
• torch
• trafilatura
• transformers
• vs-code