My goal is to create a platform that makes aerospace knowledge accessible. But what does it take to bring this vision to life? In this post, you get an inside look at how the platform is built, from features to the systems I built to make them possible.

The goal is to create a single place to find any aerospace resource you might want, with a great user experience. There are some key requirements that this leads to. First of all, it needs a lot of data that must be organized and kept up to date. The second requirement for me was that humans and agents could communicate and contribute to this open dataset.

Expanding the resources

Initially, I was adding all the data by hand. I've added hundreds of data entries manually, but this was boring and doesn't scale. So I started working on an agent that can manage the database for me. It will use the web to gather data and extract it from raw files, then organize it into our own datasets. It then also has a pipeline to schedule tasks or chronologically check that data is still up to date. This took months of programming and slow improvements, but it has gotten so good lately that it can now add hundreds of data entries autonomously, without much input from me. I think AeroVia will make it possible to create the biggest aerospace datasets by the end of 2026.

This is all managed through a dashboard I built that tracks all agents' runs built with Next.js and Convex. The most important part is the "Agent component by Convex" which is a wrapper around the AI SDK by Vercel, which lets me create autonomous agents with custom tools.

Connecting the humans and agents

An important goal of AeroVia is to connect the world and share my love for aerospace! To make this possible, I built a community page with a forum, a leaderboard, and a job-hunting page. While the forum is not very active rightnow, I am sure it will be used more once we expand our resources more and make the platform more valuable to our users. The leaderboard and points system are there to show your friends how many courses you've been reading or how many contributions you've made to AeroVia.

Let me tell you more about the contribution system because it's quite cool. Each contribution is autonomously checked by an agent to check whether it is relevant and factually correct. It can then trigger an update to the database to incorporate the contribution. This can fix everything from typos, wrong images, missing entries, and more. This makes use of the same system I told you about earlier to add to the database, but now it is triggered by feedback.

I see a future where autonomous agents like OpenClaw also contribute to expanding the datasets, but I still need to figure out exactly how this would work, as there are things to consider like quality, scaling, and security. I will surely make a blog post about that soon when I have more answers!

The mobile app is also really close to completion, and I think I'll be able to publish it this month.

List of core technologies

• Web Framework: Next.js
• Mobile Framework: Expo
• Database: Convex
• Deployment: Vercel + EAS
• User Authentication: Clerk
• File Management: Convex R2 component
• Styling: Shadcn + Tailwind
• Analytics: Posthog