Retrieval vs Cache Augmented Generation (CAG vs. RAG)
Discover the key differences between Retrieval-Augmented Generation (RAG) and Cache-Augmented Generation (CAG) in AI. Learn how RAG dynamically retrieves real-t...
6 min read
RAG
CAG
+5
Knowledge Sources
Connect and manage diverse knowledge sources—like websites, documents, and videos—to power accurate, real-time AI chatbots with FlowHunt.
What makes a general chatbot yours is the knowledge of your website and unique context. There are several ways to go about this, and many of them remain inaccessible to small businesses. FlowHunt can cater to both small and larger businesses by using the Retrieval-Augmented Generation method.
This method allows you to fully benefit from the general knowledge and capabilities of models such as ChatGPT, but with a twist. Unlike old costly methods of feeding the model training data, RAG allows you to link and upload knowledge sources to get real-time, accurate responses without the extra hassle.
What is Retrieval-Augmented Generation
Imagine you’re planning a trip to a new city and have many questions about the sights and good places to eat. You decide to ask a friend who’s an expert on the city. Instead of only relying on their memories, they quickly re-check travel guides and recent online reviews before giving you a curated answer. In other words, your friend makes sure to give you the most accurate and up-to-date answer.
This is similar to how Retrieval-Augmented Generation (RAG) works. The method seamlessly combines retrieving knowledge from an external database with the generative power of pre-trained LLM models.
Just like your friend checking the guides and reviews, your Flow will first consult your sources and only then generate an answer based on that information. This means you get more accurate and context-aware responses.
If you ask a vanilla AI about the best restaurants in the city, it will only use its training data. This data might be outdated or incomplete, leading to terrible experiences on your trip. But with RAG, the AI can pull the latest reviews and recommendations from URLs, Google, and other sources, ensuring you get the best possible advice.
Simply link and upload various types of sources with just a couple of clicks. The sources are indexed and then retrieved based on the user’s query. You control the content the chatbot uses, including information from your public website, documents you prefer not to publish, YouTube videos, and external learning materials.
Knowledge source options
The friend from our example might check various sources. They might read web pages, consult a book, check Google, or even watch a YouTube video. Your Flow can learn from all these sources, too.
But there’s a difference. While searching, your friend will make sure that the information they’re learning from is up-to-date and correct. AI can’t do that, and making sure the information is fresh and true is solely up to you.
There are three ways you can point your Flow to the right information:
Schedules
Crawl and index entire domains or single URLs periodically. Just set it and forget it. Your Flows will continue learning on their own at the scale and speed you pick.
Learn more about the Schedules feature.
Documents
Instantly access and utilize information from various document formats, HTML pages, and even YouTube videos. Either upload from a file or link a URL.
Learn more about the Documents feature.
Questions & Answer
Provide pre-defined answers to specific queries. This will ensure that the bot always stays consistent with important queries and FAQs.
Learn more about the Questions & Answers feature.
Categories of Knowledge Sources
Each new source you add needs to be categorized. These categories are completely arbitrary and meant to help you manage your growing library.
Learn more about the Categories feature.
Frequently asked questions
- What are Knowledge Sources in FlowHunt?
Knowledge Sources allow you to connect various information types—like websites, documents, and videos—to your AI chatbot, enabling real-time, accurate, and context-aware responses tailored to your business needs.
- How does Retrieval-Augmented Generation (RAG) work in FlowHunt?
RAG combines external knowledge retrieval with the generative power of AI models. Your chatbot first consults your linked sources, then generates answers based on the most up-to-date and relevant information.
- What types of knowledge sources can I link with FlowHunt?
You can link websites, upload documents, connect YouTube videos, and provide pre-defined Q&A pairs to power your AI chatbot.
- How are knowledge sources managed and organized?
Each knowledge source can be categorized for easy management, allowing you to organize your growing content library and control what your AI bot can use for responses.
- Can I automate knowledge updates in FlowHunt?
Yes, you can schedule crawls of websites or specific URLs so your chatbot’s knowledge stays fresh and relevant automatically.
Start Building Smarter AI with Your Own Knowledge
Discover how FlowHunt's Knowledge Sources feature enables you to link, categorize, and manage content for accurate, context-aware chatbot responses.
Learn more
GoogleSearch Component
FlowHunt's GoogleSearch component enhances chatbot accuracy using Retrieval-Augmented Generation (RAG) to access up-to-date knowledge from Google. Control resul...
4 min read
AI
Components
+4
Document Retriever
FlowHunt's Document Retriever enhances AI accuracy by connecting generative models to your own up-to-date documents and URLs, ensuring reliable and relevant ans...
4 min read
AI
Document Retrieval
+3