Temporal: Remember every useful thing you've read on the Web

A smarter way to save and search content on the Web.

ยท

4 min read

Temporal: Remember every useful thing you've read on the Web

Introducing Temporal

Temporal header.png

Temporal is a web app that aims to help users remember every useful thing they have read on the internet. Users just have to provide the URL to the article and Temporal magically handles the rest. It indexes the content of the article, generates vector embeddings and ensures that users have access to all of their articles at their fingertips.

Inspiration

I love it when I come across a great article on the internet. I might learn something new or discover something awesome. So I proceed to add it to my bookmarks... which is filled with several other articles. Unfortunately, that's the end of it.

A few weeks down the line I will forget that it even exists. How do I know this? Well, this happens all the time with me. Bookmarking articles on the web is messy and very difficult to get back to. There is no way to find exactly what you're looking for when you have multiple things bookmarked. You know what you're looking for and you know that it is somewhere in your bookmarks, but there's no way to find it!

Temporal solves this by indexing the actual content of a website. It also adds a layer of powerful semantic search that lets you look for concepts rather than words.

Screenshot 2022-06-30 at 6.28.34 PM.png

Walkthrough

If you want to remember everything you read on the web in a way that really works, check out Temporal!

Live website

GitHub repository

Test account credentials:

Email: test@temporal-hackathon.com

Password: linode_hashnode2022

Screenshot 2022-06-30 at 6.39.45 PM.png

Create an account ๐Ÿ‘จโ€๐Ÿ’ป

Visit Temporal and click on Login to create a new account.

Save articles on Temporal โšก๏ธ

Temporal ensures saving articles is as frictionless as possible. After logging in, open the app and click on the + button at the bottom of the website. Now all you need to do is enter the article's URL and click on Save!

Screenshot 2022-06-30 at 6.31.17 PM.png

Temporal goes ahead and indexes the contents of the website. It uses machine learning to generate a vector embedding for the document which is used for searching across all of your articles.

Search for whatever you're looking for ๐Ÿ”Ž

As you continue using Temporal to save articles, you can easily search for anything you need. Temporal uses semantic search to find articles that most closely match your query. This enables you to search for an idea instead of any specific phrase and get relevant results.

Screenshot 2022-06-30 at 6.34.02 PM.png

It's not possible to remember exact phrases from every article, especially as time passes. With Temporal, you don't have to. Semantic search makes it possible to find the information you need even if youโ€™ve forgotten where exactly it was that you saw it.

Find similar articles โœจ

When you search or save articles on Temporal, you can also find articles on similar topics in your library. Temporal does this by creating vector embeddings of your content so that it can compare how similar different things are.

Screenshot 2022-06-30 at 6.36.03 PM.png

Technologies used

Frontend

The website has been built using Next.js and deployed with Vercel.

Backend

  • Nginx has been used as a reverse proxy, ensuring secure and efficient communication between users and the server.
  • Nginx forwards the requests to a Node.js server that handles each request.
  • Communication with the database and simultaneous natural language processing are taken care of with a FastAPI server.

The three services are hosted using Linode.

Database

Temporal uses a PostgreSQL database that is hosted on a separate Linode instance.

What's next?

It has been a really exciting journey creating Temporal and participating in this hackathon. I have plans to further improve the web app and create a web extension that will allow users to save articles more effortlessly.

Long term, it would be quite interesting to support indexing audio and video content from the web as well.

Acknowledgement

I would like to thank Linode and Hashnode for organising this hackathon. I had a lot of fun participating in this and it was a great learning experience as well. I would also like to thank my friends and family for their continuous support!

ย