Inspiration: As educators & technologists, we saw how expensive and inaccessible quality IELTS preparation was for many students. We wanted to democratize access to top-tier test prep with the help of AI, with real test data, and with human mediated guidance.

What it does? IELTSY is an AI-powered IELTS coach that replaces expensive tutors. It offers structured lessons, real Cambridge-based test practices, instant feedback, and weekly expert check-ins—making IELTS prep affordable, effective, and engaging.

How we built it? Python, FastAPI, StreamLit, MCP, A2A, Perplexity, Anthropic, OpenAI, Whisper, Pillow, Pytesseract. Had initially attempted Rime + LiveKit

Challenges we ran into: Time & fitting sponsor technologies into conceptual structure.

Accomplishments that we're proud of: We unexpectedly completed the POC & submitted our work.

What we learned? This is our first hackathon. There is plenty to walk away with from this experience.
Allot time to warm-up to Git if it has been a while.
Know how DevPost works (ins-&-outs).
Don't let others (even partners/teammates) command your computer.
File handling - watch files/folders closely - unintended outcomes happen without detection. Tis better to complete work without meeting all sponsor requirements than to twist-&-bend the project to accomodate enough sponsors at the expense of incomplete/unfinished Frankensteinian work. Hours of problem-solving push is not made easier without sleep. Schedule - 'tis way funner to be early at the event.

What's next for IELTSY? We’re planning to build a large platform for SAT, TOEFL, and language learning in general.

Built With

  • a2a
  • anthropic
  • fastapi
  • mcp
  • openai
  • perplexity
  • pillow
  • pytesseract.
  • python
  • streamlit
  • streamlit-mcp
  • whisper
Share this project:

Updates