Ringchat 🤖 is a URL-powered 🔗 chatbot service.

"RingChat" is a portmanteau of "Ring" and "Chatbot". The "Ring" symbolizes the connection between the document contained within the URL link, the chatbot, and the user. In Korean, the first syllables of "Ring" and "link" are the same, allowing for a natural association. Additionally, "Ring" evokes the image of a ring, implying reliability in the answers provided by RingChat.

RingChat🤖 Key Features

1️⃣ Recommend Links along with 3 metrics

• Score
• Times attached
• Total number of bookmark

2️⃣ Customized document analysis

3️⃣ Organize scattered answers

ringchat.mp4

Service architecture

This section provides an overview of the project's core components. Each tech stack is accompanied by a discussion of its benefits, drawbacks, and the specific reasons for its inclusion in this project.

Technology	Why We Chose It	Benefits	Drawbacks/Limitations
Docker	Consistency in the development environment and rapid deployment	Resolving library dependency issues that arise in different development environments. Easy scalability during deployment.	Complexity of container image management. If not familiar with Docker, it can lead to potential delays in development time
FastAPI	High development productivity because it supports automatic documentation and asynchronous processing when developing API. Easily fulfills the essential requirements for this project	Asynchronous processing support Efficient use of resources Swagger support Easy data validation using Pydantic	Requires learning about asynchronous programming. Community size is small compared to flask
Langchain	This project aims to overcome the limitations of existing LLMs. Langchain, one of the most popular and intuitive frameworks that facilitates in-context learning, is a natural choice for achieving this goal.	It's easy to extend various functions with support for a wide range of tools. Development is relatively easy with an intuitive API.	The high level of abstraction can make fine-grained control difficult. Since it's a relatively new framework, there might be challenges in acquiring information and a higher likelihood of encountering bugs
Next.js	The only person on the team who handles frontend development is proficient in Next.js.	SSR, SSG support Improved search engine optimization Active community	For pages where data changes frequently, SSR can cause performance degradation as the server needs to render the page every time it's requested. Understanding React is necessary, which can create a barrier to entry.
Oracle 23ai	It supports Vector columns, enabling similarity search within the database. Similarity search was expected to be frequently used for URL recommendations in this project.	Direct AI model execution within the DB. AI Vector Search support. Enhanced security with SQL Firewall	Relatively high licensing fees There is a barrier to entry as it is more difficult to handle than other DBs. Relatively little reference material
OpenAI	The gpt-4o-mini model offers good performance at a relatively low credit cost. Once the charged credits are exhausted, no additional charges will occur. Code-related answers are superior compared to Gemini	Answers are less likely to be truncated in the middle Decent performance for the cost	The maximum token limit is lower compared to Gemini Multimodal processing capabilities are less developed compared to Gemini
OCI free tier	Because it provides two free VMs, each with an AMD CPU and 1GB of RAM.	Two VM instance free forever No worries about additional charges	The resources provided are inadequate for running actual services, therefore requiring an additional charge.

Documentation

Get Started

1. Clone repo
2. Set .env & .env.local files
3. Place the wallet folder
4. Run the containers

Clone repo

git clone https://github.com/Rimember/ringchat.git  # clone

Set .env & .env.local

# create .env file
cd ringchat/backend

echo "# API 
OPENAI_API_KEY = \"xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx\"

# oracle
DB_USER = \"USER_NAME\"
DB_PWD = \"PASSWORD\"
DB_DSN = \"DATA_SOURCE_NAME\"
DB_CONFIG_DIR = \"WALLET_FOLDER_PATH\"
DB_WALLET_LOCATION = \"WALLET_FOLDER_PATH\"
DB_WALLET_PWD = \"WALLET_PASSWORD\"" > .env

# create .env.local file
cd ringchat/frontend

echo "NEXT_PUBLIC_API_URL=http://backend:8000/api/v0" > .env.local 

Place the wallet folder

mv WALLET_FOLDER_PATH ringchat/backend 

Run the containers

cd ringchat 

docker compose up 

API Document

Database ERD

Contributors ✨

Thanks goes to these wonderful people

강동훈 🛠️

선경은 🛠️

이찬호 🛠️

임승원 🛠️

정지혁 🛠️