Hey Hunters! I'm Nick and today I'm launching Relevant! Relevant is my answer to being able to find interesting new content on YouTube without relying on it surfacing organically by the YouTube "algorithm". Relevant relies on a crowdsourced data model to gather information about what type of content a channel creates. Users can browse the existing corpus of information, to find everything from broad categories like "Science" and "Gaming" down to more niche one likes "Laptop Reviews". Each day we surface new categories on the "Trending" page about what new categories have popped up and what is currently a hot topic. For those who want to help contribute, you can walk through our Categorization process. For each of your subscriptions, cast your vote about what category you believe their content fits within. Once enough people have voted for a specific channel, we assign it a category and it is grouped together on the website. On a more technical note, this project has been over 4 years from start to finish. Although there was a lot of time off, a lot of those years consisted of hard work developing not only a frontend website built in NextJS but also a set of Go microservices serving APIs and running the scraping and categorisation systems. All of these systems are backed by a Postgres database, orchestrated by a Temporal service and hosted in a Kubernetes cluster. For a longer story and some more details about the website behind the scenes, check out my latest blog post! https://medium.com/@nthomson113/...

Relevant

Explore YouTube without the algorithm

Explore YouTube without the algorithm