AI Content Search (RAG) with Docs Agent | Build with Google AI
Вставка
- Опубліковано 27 лип 2024
- Learn how to build an AI-powered conversational search interface for your content using the Google Gemini API (ai.google.dev/docs). This video features Docs Agent, an open source project from Google engineers. You can extend it to let users search your content using Google's artificial intelligence models without training or tuning those models. This project uses a Retrieval Augmented Generation (RAG) AI language model pattern using a Chroma vector database and text embedding techniques for content indexing and searching. This episode includes a demo of the project, quick chat with the developers who built it, and an overview of how to extend the project to use with your content.
UPDATED: Tutorial and code project updated for Gemini API!
Resources:
Tutorial → goo.gle/3PDLD39
Code project → goo.gle/3F2Ce08
Chapters:
0:00 - Introduction
2:11 - Overview with the builders
7:10 - Extending the project
11:04 - Conclusion
In this series, Build with Google AI, explore how you can build practical solutions with Google AI technology.
Watch more episodes on Build with Google AI → goo.gle/BuildwithGoogleAI
Subscribe to Google for Developers → goo.gle/developers
#Google #AI #BuildwithGoogleAI #Gemini #GeminiAPI - Наука та технологія
Subscribe to Google for Developers for more episodes on this series, coming soon! → goo.gle/developers
Watch more episodes on Build with Google AI → goo.gle/BuildwithGoogleAI
I really like how the system provides context and links to relevant documentation sections. It makes the answers more helpful
This has absolutely game-changing implications for Content Strategists; I dream of a future where we could create content that would be structured and formatted for this kind of search optimization upfront so the "meaningful chunks" can be considered during the creation process, or reviews of current content would be made with an eye towards improving to make this possible.
I came here via the Shorts link, I'm once again absolutely amazed at what Google can do.
Thank you for watching!
The legend himself!
can this be extended to the app data residing in firestore?
The dialog flow was awesome I did something similar before 8 years with separate files, folders and sources for data
Self learning and also the bot learn and record from each user question
And then Google stopped the DF and move it to some other name the functions and ideas😊 are the same It looks like
Now I am trying to make an Ai with Google Programmable Custom Search Engine
Love this!
Thank you for watching! We’re glad you found the video helpful.
Can we extend this to chat with our PDFs ?
UPDATE: Release 0.1.3 of the Docs Agent code project adds support for PDFs! Check it out!
👍👍👍
It’s a great time for us, Google fanboys! ❤ Chrome, V8, Android, Google API, Firebase, TensorFlow, WebRTC ❤️
... and then they "sunset" them.
❤
❤❤❤ i love Google
Hello
E❤❤❤❤ EXCELLENT EXCELLENCE 🫡✌️🤞🤙🫡💘🫡 SIR 👍 AND I"M INTERESTED ✌️ SIR 👍 PLEASE GUIDE ME SIR 🫡
I don't know where to post this so I hope it's OK here. Please give us back the block feature. I don't want to see evil content or profanity when I come to youtube. Some things I am forced to see in the main fee and the shorts makes me sick to my stomach. If you just hide the channel from your channel you still see them everywhere and everything they post. I block people so I don't see their stuff, not so they can't see what I post. Please give us back the block feature. You don't monitor everything on youtube and some stuff is really perverted and disgusting and sick. Please.
Hello
Hello