MyLocal AI

AI-Powered Productivity at Your Fingertips—Secure, Efficient, and Local.

Comment

Inspiration

The idea for MyLocal AI came from the growing need for tools that could help people work more efficiently, learn faster, and interact with their content in a smarter way—all while keeping their data secure and private. We wanted to create a browser extension that combined essential productivity features like notes, flashcards, and AI-powered summaries without sacrificing data privacy. The idea was to offer a solution that works entirely on the user's device, ensuring complete control over their data.

What it does

Gradient Colorful Minimalist Coming Soon Banner (1200 x 800 px) (2) MyLocal AI is an AI-powered browser extension designed to supercharge your productivity. It provides the following key features:

  • Notes: Quickly save any information from a webpage to your personal notes.
  • Flashcards: Transform your saved notes into interactive flashcards using AI to help you memorize content.
  • Chat: Chat with the AI and ask any questions to get instant, reliable responses.
  • Summarize: Condense long articles or web pages into shorter, key-point summaries.
  • Writing: Use AI to help you draft emails, reports, or blog posts effortlessly.

All of this happens securely on your device, with no need to worry about your data being uploaded to the cloud.

How we built it

We built MyLocal AI using cutting-edge technologies to provide a seamless and fast experience:

  • JavaScript for the core logic and interactivity.
  • HTML5/CSS for the clean, intuitive interface.
  • AI Models to power the flashcards, summarization, and chat features.
  • Local Storage to ensure all data is kept securely on the user's device.
  • The extension was designed to work in a variety of modern browsers to maximize accessibility.

Challenges we ran into

One of the biggest challenges we faced was not getting the stable DX for the model. The model used to throw new error, after multiple prompt. And another challenges we have faced is setting up the dev env to start work with this thing, and I had a low specs device so that was other hectic reason for me to work with the model.

Accomplishments that we're proud of

  • User Experience: We’ve designed a clean, intuitive interface that makes using the extension simple and enjoyable.
  • AI Integration: The seamless integration of AI-powered features, like flashcards, summarization, and chat, has been a major achievement.
  • Data Privacy: We’re proud to offer a tool that puts privacy first by processing everything locally on the user's device.
  • Performance: Optimizing the extension to run smoothly without significant performance hits was another accomplishment we’re proud of.

What we learned

Through this project, we learned the importance of balancing usability with and functionality, as well as the technical challenges involved in making AI models run efficiently on the user's device. We also learned that users appreciate tools that are not only powerful but also simple to use. Continuous feedback and iteration were key to refining the user experience and improving the AI features.

What's next for MyLocal AI

Looking ahead, we plan to:

  • Expand AI capabilities by adding more advanced features, such as automatic summarization of images and videos if that's possible in future.
  • Enhance personalization by making the AI smarter and better able to understand user preferences.
  • Add collaborative features so users can share notes and flashcards with others securely.

We’re excited to continue evolving MyLocal AI to help users become more productive and work smarter every day!

Feedback for the Dev Team

  • Main Problem I have faced I have a low end device with 2GB of VRAM, so I wasn't able to run the model on that device. It was very painful for me.
  • There should be an option to run the model in CPU and RAM, If user want. I think this should be a must have feature.
  • Please If possible add a type docs for TypeScript
  • An SDK would be much nicer to work with apis,
  • And It would be very nice, If every api have the similar syntax, currently these all are different. So It's kind of killed the DX when coding.
  • Add more token capability to the summarization API, this is most important I think.

Built With

Share this project:

Updates