Use Algolia MCP Server & VSCode GitHub Copilot to Find and Submit Unique Projects in Dev Challenges! ❤️‍🔥

This is a submission for the Algolia MCP Server Challenge

Introduction

Recently I've found myself participating in quite some Dev Challenges. I'm Loving it!

However, the problem is: Most of my Dev Challenge submissions are done on the weekends. That means by then, there are already many submissions.

That's great! But what's the point in working on something someone had already done! I mean, it's not always a bad thing, but wouldn't you rather work on something unique!

In comes Algolia MCP Server Challenge!

So I thought, would the Algolia MCP Server make it easy for me to check if what I'm going to work on is at least unique so far?

This is what this project is all about.

What I Built and How I Utilized the Algolia MCP Server

• Installed Algolia MCP Server on VSCode with GitHub Copilot (Linux).
• Developed a CLI Tool to Fetch dev.to Challenge Submissions using the Forem API.
• Developed a Web Interface to check the Submissions locally.
• Developed an MCP (Model Context Protocol) Server named dcs to get the Submission data in GitHub Copilot Agent Mode.
• Then indexed that Challenge submission data in Algolia, using Algolia MCP Server, by simply asking:

#dcs fetch algoliachallenge tag submissions, get all the submissions from this tag, and add them to my Dev Challenges app in #algolia

• Finally, asked GitHub Copilot in Agent mode if my project idea is unique!

Once I built everything, all I had to do was ask:

#algolia in Algolia Challenge did anyone submit something related to GitHub Copilot?

And there it was!

• I didn't have to dig deep into all the submissions on dev.to
• Didn't have to do any complex search

Just one single command to Algolia MCP server and the AI Agent took care of the rest!

In my mind, I've already won something from this challenge! 🚀

I'm sure other developers are going to love this technique, and future Dev Challenges are going to receive more unique submissions using similar solutions.

Who knows, may be dev.to will provide their own MCP server in the near future! 💫

Demo

The following GitHub repo has all the necessary code and instructions for:

• Using GitHub copilot with Algolia MCP Server and DCS (Dev Challenge Submission) MCP Server
• DCS CLI
• DCS Web App

DCS: Dev Challenge Submission - GitHub Repo

Key Takeaways

• I've made the initial setup using vite and pnpm.
• Then mostly vibe coded the solutions using GitHub Copilot in Agent Mode with Claude Sonnet 4. But I did have to do some work when Claude started to make mistakes. So it's not all vibe coded work.
• Algolia MCP Server documentation doesn't provide instructions for Linux or GitHub Copilot, but with a bit of tinkering, I could make it work.

.vscode/mcp.json file:

{ "servers": { "algolia": { "command": "${input:algolia_mcp_dir}/algolia-mcp" }, "dcs": { "command": "${input:dcs_mcp_dir}/server" } }, "inputs": [ { "id": "algolia_mcp_dir", "type": "promptString", "description": "Path to the Algolia MCP directory", "password": false }, { "id": "dcs_mcp_dir", "type": "promptString", "description": "Path to the dev.to Challenges MCP directory", "password": false } ] } 

• GitHub Copilot's support for Algolia MCP Server is quite good!
• Although, only Claude models support Algolia MCP Server in GitHub Copilot, GPT models don't support it! Gives 400 error.
• That said, GitHub's own MCP Server works with GPT 4.1 in GitHub Copilot Agent Mode. Didn't have time to debug why Algolia MCP Server only works with Claude models. Perhaps Algolia team will work on this?

Overall, participating in this challenge was really a happy experience for me.

I hope other developers will also be benefited from this project. ❤️