Until yesterday, our goal clarification flow was still running on a hardcoded logic — recognizing goal keywords like “app” or “coding” and mapping them to fixed follow-up questions. I knew this had to change.

So I asked for a real connection to OpenAI — and today, I finally made it happen. The full chain is now working: GitHub → Vercel → OpenAI. The API endpoint is live and receiving requests dynamically.

The app itself isn’t ready to handle the response yet — which means I’m getting errors on the front end. But seeing that the OpenAI API was charged confirms one thing: Microstep really talked to OpenAI. That tiny $0.02 charge meant everything.

Tomorrow, I’ll work on parsing the API output and displaying the questions and options inside the app UI. Excited for what comes next.

Day 5: OpenAI API 연동 디버깅의 날

어제까지는 하드코딩으로 목표를 식별해서 세부화 시키는걸로 코딩돼있던걸 발견해서 openai 를 연결해달라고 했더니 이제야 GitHub - vercel - openAI 연결고리가 완성되었다. 아직 앱에서 openAI에서 출력받은 값을 처리할 수 있는 상태가 아니라서 앱에서는 에러가 뜨고있지만 API 과금된걸 보면 Microstep 이 API 를 활용했다는걸 확인했다. 정말 값진 2cent 였다! 내일부터는 openAI 에서 받은 출력값이 앱에서 display 되도록하는 작업을 진행할 계획이다.