Building a Fortune Teller App with OpenAI From Chatbot to Voice Assistant

Building a Fortune Teller App with OpenAI From Chatbot to Voice Assistant

Transcript

1. Building a Fortune Teller App with OpenAI: From Chatbot to Voice Assistant
2. • Introduction • OpenAI model • OpenAI API • LINE x OpenAI • Conclusion Agenda
3. Introduction
4. AI is All Around Us!
5. AI is All Around Us!!! Source: https://www.pcgamer.com
6. We’re Building an AI Fortune Teller App! Today, we’re going back to the basics of OpenAI to create something special for Thai people: an AI Fortune Teller App the ultimate use case!
7. Demo 🔮
9. How do we build this application? 🔍
10. OpenAI Models
11. What is AI Model? Input Output Model “An AI model is a program that learns from data to perform tasks like understanding text, recognizing images, or generating speech based on the patterns found in the data.” 
 “Input as Text” 
 “Output as Text” Prompt
12. What is Multimodal Model? Input Output Multimodal
 Model “A multimodal model 
 is a type of model that handles multiple data types, like text, images, 
 or audio.” 
 “Input as Text, Image 
 or Audio” 
 “Output as Text, Image 
 or Audio”
13. OpenAI Model Group Model Speech 
 Generation Image 
 Generation Text 
 Generation
14. OpenAI Flagship Models Input Output Model GPT-4o, GPT-o1, etc Whisper TTS DALL·E
15. OpenAI APIs
16. OpenAI API (OpenAI architecture) HTTP
 Request HTTP
 Response Model OpenAI API Simple HTTP Protocol
17. Text Generation API
18. Completion API Request Response Model
 GPT-4o OpenAI API Looks good, right? But remember, this API is stateless! (Text Generation API)
19. Assistant API (Text Generation API) Stateful API with many more features!!! Request Response
 Model
 GPT-4o OpenAI API
20. Speech to Text API
21. Audio Transcription API (Speech to Text API) Request Response
 Model
 Whisper OpenAI API
22. Text to Speech API
23. Audio Speech API (Text to Speech API) OpenAI API Request Response
 Model
 TTS OpenAI API
24. Realtime API
25. Realtime API (Speech to Speech API) OpenAI API Request Response
 Model
 GPT-4o-realtime OpenAI API
26. Playground
27. LINE X OpenAI Let’s building a Fortune Teller App with LINE and OpenAI
28. Architecture of Fortune Teller App (Chat Assistant Version) LINE Messaging API Webhook
29. Architecture of Fortune Teller App (Voice Assistant Version) LIFF
30. Architecture of Fortune Teller App (Voice Assistant Version with Realtime API) LIFF
31. Use case More fl exibility Better for complex conversation Low latency Less complexity Each option offers its own benefits! &
32. Conclusion
33. LINE x OpenAI = Future! 🔮