LearnSync – A Unified Knowledge Explorer
This is a submission for the Algolia MCP Server Challenge
What I Built
LearnSync is a unified search engine for developers that aggregates and ranks learning content from three major platforms:
- DEV.to articles
- GitHub repositories
- YouTube tutorials
Whether you’re diving into a new framework or exploring a deep-dive into Web3, LearnSync brings the best learning materials across sources into one clean, fast, searchable interface.
Demo
Live App: learnsyncsub.netlify.app
GitHub Repository: https://github.com/pulkitgovrani/LearnSync
Video Walkthrough:
How I Utilized the Algolia MCP Server
I leveraged the Algolia MCP Server to power the unified search layer across content types.
Content Ingestion Logic:
-
DEV.to: I used their public
/articlesendpoint, supporting filters liketag,username,page, andstateto fetch fresh and relevant articles.
Example structure:
const url = 'https://dev.to/api/articles?tag=react&page=1&per_page=20&state=fresh';
-
GitHub: I utilized the
https://api.github.com/search/repositoriesendpoint to fetch trending repositories around a topic, sorted by stars and updated date.
const url = `https://api.github.com/search/repositories?q=react&sort=stars&order=desc&page=1`;
-
YouTube: I pulled topic-based tutorials from the
youtube/v3/searchendpoint using the YouTube Data API, filtering for recent and relevant videos.
const url = `https://www.googleapis.com/youtube/v3/search?part=snippet&q=react tutorial&type=video`;
All fetched content was normalized to a common schema and indexed into a single Algolia index (learnsync). The frontend uses this index to perform instant searches with filtering and relevance ranking powered by Algolia.
Key Takeaways
- Normalizing article, repo, and video data into a unified format was challenging but crucial.
- Algolia made building a lightning-fast search UI incredibly easy and scalable.
- Writing a backend that ingests, formats, and syncs data from three APIs taught me a lot about rate limiting, data structure design, and batching.
- The project evolved from a basic search bar to a continuously updated, API-powered learning tool.
Team
Built solo by @pulkitgovrani
