St.ChatInterface

A Streamlit powered chat interface to use powerful LLM models by OpenAI and Anthropic.

Features

• Cheaper than individual ChatGPT and Anthropic subscriptions.
• Saves all the chats and their respective responses in a local database.
• Data will not be used to train LLMs.
• Allows function calling.
• [Coming Soon] Allows model to produce interactive elements (graphs, maps, etc).
• [Coming Soon] Multimodal model support.

Models Available

Model Name	Price per 1M Input Tokens	Price per 1M Output Tokens	Multimodal
gpt-3.5-turbo-0125	$0.50	$1.50	❌
gpt-4-0613	$30.00	$60.00	❌
gpt-4-0125-preview	$30.00	$60.00	❌
gpt-4-1106-vision-preview	$10.00	$30.00	✅
claude-3-haiku-20240307	$0.25	$1.25	✅
claude-3-sonnet-20240229	$3.00	$15.00	✅
claude-3-opus-20240229	$15.00	$75.00	✅

Installation

Python 3.7+ is required to run this project.

git clone https://github.com/saksham2001/St.ChatInterface
cd St.ChatInterface
pip install -r requirements.txt

Setup

Setup the database to save all the chats and their respective responses.

mkdir data
python3 create_db.py

Usage

Replace <your_api_key> with respective API keys for your OpenAI and Anthropic accounts.

Note: You only need to set OPENAI_API_KEY if you are using OpenAI's models and ANTHROPIC_API_KEY if you are using Anthropic's models. If you want to use both, you need to set both.

export OPENAI_API_KEY='<your_api_key>'
export ANTHROPIC_API_KEY='<your_api_key>'

streamlit run app.py

[Optional] To give access to weather information to the models, you need to set OPENWEATHER_API_KEY environment variable.

export OPENWEATHER_API_KEY='<your_api_key>'