AI Study Assistant

Overview

This project is a lightweight AI-powered study assistant built with FastAPI and the OpenAI API.
It provides endpoints for text summarisation, structured note generation, and Q&A, making it easier to work with study materials.

Features

• /summarise: Produce a clear, concise summary of input text
• /notes: Turn raw text into structured study notes with headings and bullet points
• /qa: Answer questions, optionally grounded in a given context passage

Tech Stack

• Backend: FastAPI (Python)
• AI: OpenAI gpt-4o-mini via the async client (AsyncOpenAI)
• Data modelling: Pydantic request/response models

How It Works

1. The client sends a JSON request to one of the endpoints.
2. The API constructs a system prompt and user message based on the endpoint.
3. The service calls the OpenAI Chat Completions API through a shared helper function.
4. The response is validated and returned as a structured JSON payload.

The OpenAI client is lazily initialised using the OPENAI_API_KEY environment variable.
If the key is missing, the API responds with a 503 error so configuration issues surface early.

Endpoints

• POST /summarise
• POST /notes
• POST /qa

You can explore and test them via the built-in Swagger UI at /docs once the server is running.

Possible Future Improvements

• Add authentication and basic rate limiting
• Add a simple frontend (React or Next.js) to interact with the assistant
• Introduce retrieval-augmented generation (RAG) using a vector store
• Log prompts and responses for analysis and prompt optimisation

Author

Jingjing Wang