Nothing Special   »   [go: up one dir, main page]

Skip to content

josancamon19/openai-realtime-console

 
 

Repository files navigation

Interactive Learning Console

This is a customized and improved version of the openai realtime dashboard for learning.

What's different?

  • Long sessions are supported, on original dashboard, 15 mins is the max.
  • Conversation history is maintained in the context of the whole conversation, thus every time you connect to a topic, the model has the context of the whole conversation.
  • A right side panel generates a mermaid graph of the concepts learned.
  • Follow up questions are generated to further the learning.
  • Tavily is included as a tool call for web search, sometimes wanted to learn about latest research about a topic.


Image 1

Image 3

Image 2

Table of Contents

Getting Started

npm install

Usage

  1. Start a Session: From the HomePage, create a new topic or continue from an existing one.
  2. Set-up api keys: In the SettingsPage, set-up your OpenAI and Tavily api keys.
  3. Interact in Console: In ConsolePage, view the conversation log, generate follow-up questions, and review the Mermaid.js graph summarizing concepts.
  4. Audio Visualization: View audio waveforms for both user and assistant audio streams.
  5. Mermaid Graph Generation: As you converse, a visual summary is generated automatically.

Buttons and Controls

  • Connect/Disconnect: Manages the real-time session state.
  • Options Menu: Provides access to ChatGPT for review, graph regeneration, and transcript copy.
  • Follow-Up Questions: Generated by AI to further conversation.

Dependencies

  • React: Main library for creating UI components.
  • Framer Motion: For smooth animations.
  • Mermaid.js: For graph generation.
  • OpenAI API: Integrates real-time audio and text responses.
  • Tavily: For external data search in conversations.

About

Better version of openai-realtime dashboard but for learning.

Topics

Resources

License

Stars

Watchers

Forks

Languages

  • TypeScript 53.1%
  • JavaScript 43.0%
  • SCSS 2.1%
  • HTML 1.4%
  • CSS 0.4%