Project Instructions

Build a chatbot that summarizes a given subject prompt for a specified audience age group, providing information that is appropriate and digestible for that age.

Tech Stack

Backend: Python FastAPI
Database: Vector database (for RAG)
LLM Integration: LangChain, OpenAI (RAG, chunking, function-calling)

Functional Requirements

Accept user input for:
- Audience age group (user must select from options such as: Early Elementary [5-8], Late Elementary [9-11], Middle School [12-14], High School [15-18])
- Subject prompt (text)
If the user submits a question without selecting an age group, respond with a friendly message reminding the user to first select an age group.
Summarize and explain the subject prompt in natural language appropriate for the selected school-age group.
Use RAG (Retrieval Augmented Generation) to enhance responses with relevant, up-to-date information.
Store and retrieve context chunks in a vector database.
Support function-calling with parameters:
- ```
context
```
  : Age group
- ```
explain
```
  : Text to summarize
Expose API endpoints for:
- Submitting a prompt and age group

Implementation Notes

Use LangChain for chunking, retrieval, and LLM orchestration.
Use OpenAI or similar LLM for summarization.
Ensure summaries are tailored to the cognitive level of the specified age group.
Design the API to be easily consumable by the Vue.js frontend .

Recommended Sources for Ingestion

The chatbot should ingest articles and information from the following reputable sources:

These sources are reputable, frequently updated, and cover a range of general knowledge and educational topics suitable for summarization for different age groups.

Project Structure Recommendations

Organize by Service/Responsibility

Separate frontend and backend code into distinct directories
Use a modular structure for both frontend and backend
Example structure:

project-root/
├── frontend/                  # Vue.js app for user interface
│   └── src/
│       ├── components/        # Reusable UI components (AgeSelector, PromptInput, SummaryDisplay)
│       ├── views/             # Main views/pages of the app
│       └── services/          # API client for backend communication
│   └── public/                # Static assets (favicon, index.html)
│   └── package.json           # Frontend dependencies and scripts
├── backend/                   # FastAPI app for API and business logic
│   ├── api/                   # API route definitions (prompt submission, summary retrieval)
│   ├── services/              # Business logic (summarization, RAG, chunking)
│   ├── models/                # Pydantic models and DB schemas
│   ├── vectorstore/           # Vector DB integration and management
│   └── main.py                # FastAPI entry point
│   └── requirements.txt       # Backend dependencies
├── tests/                     # Unit and integration tests for backend and frontend
├── .env                       # Environment variables (not committed)
├── README.md                  # Project overview and setup instructions
└── .gitignore                 # Files and directories to ignore in version control

Use Environment Configuration and Version Control
- Store secrets (API keys, DB URIs) in
```
.env
```
  files (never commit to git)
- Use
```
requirements.txt
```
  for Python dependencies
- Use
```
package.json
```
  for frontend dependencies
- Add a root-level
```
README.md
```
  with setup and usage instructions
- Use
```
.gitignore
```
  to exclude node_modules, pycache, .env, etc.
Modularize RAG and LLM Integration
- Place all LangChain, OpenAI, and vector DB logic in a dedicated backend module (e.g.,
```
/app/services/rag.py
```
  )
- Abstract summarization and retrieval logic behind service interfaces
- Keep API endpoints thin—only handle request/response, delegate logic to services
- Write clear docstrings and type annotations for maintainability

Functional Requirements

Accept user input for:

Audience age group (user must select from options such as: Early Elementary [5-8], Late Elementary [9-11], Middle School [12-14], High School [15-18])
Subject prompt (text)

If the user submits a question without selecting an age group, respond with a friendly message reminding the user to first select an age group.

Summarize and explain the subject prompt in natural language appropriate for the selected school-age group.

Use RAG (Retrieval Augmented Generation) to enhance responses with relevant, up-to-date information.

Store and retrieve context chunks in a vector database.

Support function-calling with parameters:

```
context
```
: Age group
```
explain
```
: Text to summarize

Expose API endpoints for:

Submitting a prompt and age group

Project Structure Recommendations

Organize by Service/Responsibility

Separate frontend and backend code into distinct directories
Use a modular structure for both frontend and backend
Example structure:

project-root/
├── frontend/                  # Vue.js app for user interface
│   └── src/
│       ├── components/        # Reusable UI components (AgeSelector, PromptInput, SummaryDisplay)
│       ├── views/             # Main views/pages of the app
│       └── services/          # API client for backend communication
│   └── public/                # Static assets (favicon, index.html)
│   └── package.json           # Frontend dependencies and scripts
├── backend/                   # FastAPI app for API and business logic
│   ├── api/                   # API route definitions (prompt submission, summary retrieval)
│   ├── services/              # Business logic (summarization, RAG, chunking)
│   ├── models/                # Pydantic models and DB schemas
│   ├── vectorstore/           # Vector DB integration and management
│   └── main.py                # FastAPI entry point
│   └── requirements.txt       # Backend dependencies
├── tests/                     # Unit and integration tests for backend and frontend
├── .env                       # Environment variables (not committed)
├── README.md                  # Project overview and setup instructions
└── .gitignore                 # Files and directories to ignore in version control

Use Environment Configuration and Version Control

Store secrets (API keys, DB URIs) in
```
.env
```
files (never commit to git)
Use
```
requirements.txt
```
for Python dependencies
Use
```
package.json
```
for frontend dependencies
Add a root-level
```
README.md
```
with setup and usage instructions
Use
```
.gitignore
```
to exclude node_modules, pycache, .env, etc.

Modularize RAG and LLM Integration

Place all LangChain, OpenAI, and vector DB logic in a dedicated backend module (e.g.,
```
/app/services/rag.py
```
)
Abstract summarization and retrieval logic behind service interfaces
Keep API endpoints thin—only handle request/response, delegate logic to services
Write clear docstrings and type annotations for maintainability

<!-- Use this file to provide workspace-specific custom instructions to Copilot. For more details, visit https://code.visualstudio.com/docs/copilot/copilot-customization#_use-a-githubcopilotinstructio

Project Instructions

Tech Stack

Functional Requirements

Implementation Notes

Recommended Sources for Ingestion

Project Structure Recommendations

Related Skills

Nano Banana Pro

Markdown Converter

1password

Project Instructions

Tech Stack

Functional Requirements

Implementation Notes

Recommended Sources for Ingestion

Project Structure Recommendations