Session 1 · First prototype
First prototype
This is what we did in Session 1: make the Open Circuit repository your own, run it once end to end, and edit the code yourself.
You won't build a new MCP server here. This is where you first experience the using side, with tools that are already connected. The building side comes in later sessions.
1. Run it once end to end
Start the practice web app and go from a scene to a result once. It takes about 10 minutes. → One full round
Up to here, you have used something someone else made, as it is.
2. Make the repository your own
The folder you have open now is the tool folder the install script downloaded. If you edit it, your changes will be overwritten at the next update. So fork it first.
- Fork the Open Circuit repository on GitHub. → Forking
- Download it to your computer. → Cloning to your computer
- Open that folder in Cursor.
Now you can change whatever you like. It is disconnected from the original.
3. Edit it with Cursor
Open the chat (Ctrl/⌘+L) and tell it what to do in plain words. You don't need to find the files yourself. Just write what you want to change, and the AI will find the right place.
실습 웹의 얼굴 필터에 하나를 더 추가하고 싶어.
지금 어떤 필터들이 어디에 정의돼 있는지 먼저 알려 줘.
결과가 쌓이는 격자를 한 줄에 두 칸씩 크게 보이도록 바꿔 줘.
문장 칸에 아무것도 없을 때 나오는 안내 문구를 내 말투로 바꿔 줘.
(In English, these prompts read: "I want to add one more face filter to the practice web app. First tell me which filters exist now and where they are defined." / "Change the results grid so it shows two large cells per row." / "Rewrite the hint text shown when the sentence field is empty in my own voice." You can write your prompts in English.)
After editing, restart the practice web app and check. How to start it is in Reopen.
4. Push it to GitHub
Commit your changes and push them. → Commit and push
Once you get here, there is one continuous line from setting up your tools to your own edited code being online. That was the goal of Session 1.
Going further
- What if you want to publish what you made on the internet? → Publishing on the internet
- How do you bring updates from the original repository into your fork? → When the original is updated
- What if you want Cursor to make images directly, without going through the practice web app? → Why an API?