An advanced AI assistant that processes both audio and video inputs to provide interactive responses, utilizing LLMs like LLama3.1 using groq api, gemini-1.5-flash.
We Optimized LangChain-based inference pipeline with message history management, enabling accurate voice command recognition and dynamic interactions, even in noisy environments.
We need an OPENAI_API_KEY or groq api and a GOOGLE_API_KEY to run this code. And then we have to store them in a .env file in the root directory of the project.
Create a virtual environment with conda or usual python, update pip, and install the required packages:
Virtual Environment Using pip
$ python3 -m venv .venv
$ source .venv/bin/activate
$ pip install -U pip
$ pip install -r requirements.txt
Virtual Environment Using conda
$ conda create -n venv python=3.11
$ conda activate venv
$ pip install -r requirements.txt
Run the assistant:
$ python assistant.py