jchung7v / ask-doctor-brown Goto Github PK

Ask Doctor Brown

This is a web application utilizing Next.js and Tailwind CSS, hosted securely on AWS.

See My Application · Report Bug

Software Development Plan for Child-AI Verbal Interaction Web Application

• Background

  • Any parent with a young child knows that once a child begins to speak, the questions never cease. While some questions can be answered off the top of a parent’s head, others may be more challenging and require some online searching.
  • Given that we are in 2023, there's a more efficient way to find answers than traditional searching – that is, through AI. Even young children can query an AI and receive almost immediate answers. However, a challenge arises because many of these young children aren't familiar with computers and can't type yet.
  • This issue can be resolved by integrating voice recognition at the front end and text-to-speech technology at the backend, encouraging direct child-AI interactions regardless of the child's literacy level.

• Target Audience

  • Children aged between 5 and 8, who may find it challenging to type but can verbally pose questions.

• Goal

  • To develop a web application that facilitates verbal interaction between children and AI, particularly ChatGPT.

• Technical Specifications:

  • Frontend: Utilizing Next.js with styling from Tailwind CSS.
  • Backend: Serverless architecture using AWS Lambda, with two functions: One connected to OpenAI Server and the other to AWS Polly Server.
  • API: Managed and routed through AWS API Gateway.
  • Hosting: Deployed on AWS Amplify.
  • Database: No database is utilized for this project.

• Timeline

  • Two weeks until the completion of prototype development.

• Budget

  • 10 Canadian dollars for API usage during the prototype phase.

• Iteration Plan

  • No further iterations until the prototype reaches completion.

• Features & Priority score

  • Voice recognition to transcribe user inputs on the screen, so they can visualize their questions (10)
  • AI-generated responses converted into audible, human-like speech (10)
  • Send user queries to AI for answers (10)
  • Implement a filter to exclude any violent or inappropriate content from the answers (9)
  • Provide start, stop, reset, and submit buttons to enhance user control (8)
  • Provide an intuitive and child-friendly instruction guide (5)
  • Allow users to enter their names before initiating the application (7)

• Features Excluded from the Prototype

  • User login (5)
  • User profile creation and management (5)
  • Archive all interactions between the user and AI for easy future access (8)
  • Allow users to bookmark preferred answers (5)

• Feedback (Last Updated Aug 16, 2023)

  • Allow users the option to type in questions. This will also enable corrections for potential transcription errors from voice recognition.
  • (Applied) Name change: The character is now referred to as Doctor Moustache instead of Doctor Brown.
  • Merge the “start” and “stop” functionalities into a single button that toggles based on its current state.
  • (Applied) Include an image of Doctor Moustache on the welcome page for user engagement.
  • (Applied) Color code the microphone icon - green when active and red when deactivated.
  • (Applied) Ensure that the 'reset' function clears all fields, including the answer.

• Screenshots