What I built
A suggestion and search engine for dev.to
Category Submission:
Wacky Wildcards
App Link
Screenshots
Description
A search engine and suggestion system for articles published on dev.to. Discover with ease related content of dev.to, by searching for topics of interest or by finding articles that relate to a particular one.
For example if you are reading an article on dev. to with this URL:
https://dev.to/nikkilopez2/understanding-vector-databases-49n3
you can change dev.to
to devtoo.dev
like this:
https://devtoo.dev/nikkilopez2/understanding-vector-databases-49n3
and discover more related content.
Link to Source Code
https://github.com/tzador/devtoo
Permissive License
It is released under MIT License
Background (What made you decide to build this particular app? What inspired you?)
When reading articles on dev.to, only 4 related articles are suggested at the end. I wanted a suggestion and search engine that would allow to easily discover related content on dev.to.
How I built it (How did you utilise GitHub Actions or GitHub Codespaces? Did you learn something new along the way? Pick up a new skill?)
GitHub actions where utilised to fetch the fresh content from dev.to by scanning the latest sitemaps. They are run using CROM scheduler and update the database with new content.
For database I used for the first time PostgreSQL with pgvector
extension, which indexes OpenAI based vector embeddings based on Title, Description and the Tags of articles. This allows to find related articles to a given one in a speedy manner.
Additional Resources/Info
Additionally a chrome extension and a bookmarklet has been build, so that with on click you can jump from dev.to article to a list of related ones. They also work on other websites that provide OpenGraph metadata, which can be vectorised and used as a query in our system.
Top comments (5)
Hey, cool project!.
May i know how did you host your database?
I used PostgreSQL provided by vercel, with pgvector extension, to be able to query the vector embeddings for search and related articles queries.
Thanks for your reply
Coooool
Very cool project 🌱🙌🏼