Using Voiceflow with Mistral AI or Open Sourced AI Model

Описание к видео Using Voiceflow with Mistral AI or Open Sourced AI Model

Did you know you can use Voiceflow's Knowledge Base as a Vector Database with Open Sources Models like Mistral AI or 3rd party models like Perplexity?

This tutorial goes over how to use our new functions feature to search the Voiceflow Vector Database and send relevant chunks to a 3rd party model like Mistral AI.

Voiceflow lets you build complex AI Assistants faster.

You can download the two functions used in the video here
Extract Chunks: https://www.voiceflow.com/function/ex...
Send Query to Mistral 7xB: https://www.voiceflow.com/function/se...

Learn more about functions here:    • Functions: create reusable custom log...  

🎯 Key Takeaways for quick navigation:

00:00 🧠 Introduction to using Voiceflow with Mistral AI or Open Sourced AI Model
- Introduction to Mistral AI, an open-source AI model.
- Mention of tools like Perplexity building on top of Mistral AI.
- Explanation of using Voiceflow's knowledge base as a vector database with third-party large language models.
00:55 🛠️ Setting up Voiceflow's Knowledge Base and Data Sources
- Instructions on setting up Voiceflow's knowledge base.
- Uploading data sources and the process of vectorizing them.
- Demonstrating how chunks are processed and organized in the knowledge base.
03:15 🤖 Using Voiceflow Knowledge Base API and Custom Large Language Models
- Explanation of using the Voiceflow knowledge base API to send questions and retrieve chunks.
- Mention of similarity scores for chunk relevance.
- Introduction to using custom large language models like Mistral AI with Voiceflow.
05:08 📦 Leveraging Functions for Advanced API Calls
- Introduction to using functions in Voiceflow for complex API calls.
- Explanation of input and output variables in functions.
- Demonstrating how to create a function for making API calls to models like Mistral AI.
08:48 🌐 Integrating Mistral AI with Voiceflow for Custom Responses
- Connecting Voiceflow's knowledge base and Mistral AI using functions.
- Passing questions and chunks to Mistral AI for custom responses.
- Explanation of the system prompt and answer retrieval from the Mistral AI API.
15:11 🤖 Testing the Integrated System and Final Thoughts
- Testing the integrated system in a chatbot.
- Demonstration of a custom bot responding with information from Mistral AI.
- Conclusion and encouragement to experiment with open source models and Voiceflow.


***
The collaborative platform to build AI agents. Use Voiceflow to design, test, and launch chat or voice AI agents — together, faster, at scale.

Join our Discord community
👾   / discord  

Kickstart your next project with our templates
🚀 https://www.voiceflow.com/templates

Our Links 🔗
👉 Start building today: https://www.voiceflow.com
👉 Subscribe: https://bit.ly/3am22nf
👉 Twitter: https://bit.ly/2xrXZqV
👉 LinkedIn:   / voiceflowhq  
👉 Publication: https://www.voiceflow.com/blog

Комментарии

Информация по комментариям в разработке