Inspiration

Learning something new can be difficult when you don't know where to start. There are thousands of tutorials, videos, and articles online, but learners still have to figure out what to learn first, what comes next, and which resources actually fit their goal.

We wanted to build something that makes that process easier.

LearnForge lets someone simply say what they want to learn and turns that goal into a step-by-step learning experience. We also wanted to make learning more visual, so LearnForge can generate video lessons with explanations.

The project was also a chance for us to explore what it feels like to build a real product with AI agents handling different parts of the development process.

What it does

LearnForge AI helps anyone learn a skill or topic step by step.

A user enters something they want to learn, such as "I want to learn Python," and LearnForge creates a structured learning path.

The learner can then go through the steps, get explanations, and generate visual or video lessons for the topics they are learning.

It is designed for anyone who wants to learn — students, beginners, professionals, creators, career switchers, or anyone curious about a new skill.

How we built it

We built LearnForge as a web application and used Agent Orchestrator as our coding workspace.

AO helped us break the work into tasks, plan features, delegate development work to agents, review changes, test the application, and move the project toward a working release.

For the application itself, we used React, Next.js, TypeScript, Tailwind CSS, and AI services for generating learning paths, explanations, and visual learning content.

The main idea was to use agents throughout the development process instead of treating AI as only a feature inside the final product.

Challenges we ran into

One challenge was making the AI-generated learning path feel useful instead of looking like a normal AI response.

We had to improve how the learning goal was processed and how the generated content was structured.

Another challenge was getting the different parts of the product to work together, especially the learning path, explanations, and video generation.

Using multiple agents also meant we had to review their work, test changes, and make sure everything still worked together.

Accomplishments that we're proud of

We are proud that LearnForge became a working product instead of staying as an idea.

A learner can enter a goal and get a structured learning experience, with explanations and visual/video content.

We are also proud of using Agent Orchestrator throughout the development process. It helped us experience a different way of building software where agents can help with planning, coding, reviewing, and testing.

The final result is a product that we can actually demonstrate and continue improving.

What we learned

We learned that using AI agents for development is not just about asking an AI to write code.

Good planning, clear tasks, reviewing changes, testing, and communication between different pieces of work are still important.

We also learned that building an AI-powered product requires a lot of attention to how the AI output is structured and how the user experiences that output.

What's next for LearnForge AI

We want LearnForge to become a complete AI learning companion.

Next, we want to add progress tracking, adaptive learning, quizzes, exercises, personalized revision, learning reminders, and more interactive learning experiences.

We also want to improve the generated video lessons and make them more useful for different types of learners.

Our long-term goal is to make LearnForge a place where anyone can come with a learning goal and have AI guide them from the first step to confidently using the skill.