Easy-RAG

Current Logo
Database Connection
About Page
ChatBot Interface
Software embedding on a 3rd party website

Inspiration

The idea for EasyRAG was born out of the need to overcome the limitations of existing chatbot solutions. We noticed that while many platforms offer chatbot creation, they often come with restrictions, such as rate limits, high costs, and lack of customization. Our inspiration was to democratize access to powerful AI tools and allow users to harness the full potential of Large Language Models (LLMs) by giving them the ability to train custom chatbots with their own data, all while maintaining ease of use and affordability.

What it does

EasyRAG is a versatile platform that allows users to create custom chatbots by integrating their own data and leveraging a variety of LLMs. Users can upload their documents, train a chatbot for free, and even embed it directly into their websites using our SDK. The platform offers advanced features like inserting or deleting context on the fly, making it a powerful tool for dynamic and adaptive chatbot interactions. It’s designed to work seamlessly with a wide range of input formats and provides full control over the chatbot’s training process. You can even connect your own database.

How we built it

We built EasyRAG using a robust tech stack that includes TypeScript, AWS, NextJS, Python, and Convex for the backend. Our platform utilizes state-of-the-art embedding models such as text-embedding-ada-002 and all-MiniLM-L12-v2 to convert text into vector embeddings. These embeddings are then fed into the user’s LLM of choice to generate responses. The entire process is designed to be intuitive and user-friendly, ensuring that even users with limited technical knowledge can easily create and deploy custom chatbots. suring that our platform remains both flexible and powerful enough to meet diverse user needs.

What's next for EasyRAG

Moving forward, we plan to enhance EasyRAG by expanding its compatibility with more LLMs and further refining the user experience. We’re also looking to add new features that will allow users to visualize chatbot performance metrics and fine-tune their bots more effectively. Recent papers have envisioned the use of TAG(Table Augmented Generation) which further reduces latency and improves throughput. Here's a few other things we are thinking about:

Develop a Website Chatbot: Design and implement a chatbot specifically for websites to enhance user interaction and showcase projects effectively-want recruiters to ask your chatbot questions about your portfolio? We got a chatBot for you.

Create and Optimize Database Column Embeddings: Develop and store embeddings for database columns, ensuring compatibility with databases beyond PostgreSQL, to improve data retrieval and processing.

Automate Website Scraping for Chatbot Integration: Streamline the process where clients provide a URL, enabling us to scrape the website content and integrate a customized chatbot seamlessly.