LeIA is an application that leverages artificial intelligence models for audio and video transcription. Users can transcribe new files or consult previously transcribed cases.
⏳Implementing the FastAPI application and routes at this moment. Please wait a fell days.
- Repository: https://github.com/andrecorumba/leia
- Documentation: https://andrecorumba.github.io/leia/
- Web Version for Testing: https://andrecorumba-leia-appapp-web-b757de.streamlit.app
The project is implemented in Python version 3.12.2.
streamlit run streamlit_app/streamlit_main.py
os: A library for interacting with the operating system, enabling the manipulation of file paths, directories, environment variables, etc.whisper: A library for audio and video transcription (https://github.com/openai/whisper).pandas: A library for working with tabular data, supporting manipulation, cleaning, analysis, and visualization.sqlite3: A library for working with SQLite databases, a widely used embedded relational database.streamlit: A library for creating interactive web applications for data analysis and visualization, allowing users to build interactive data analysis dashboards and control panels.streamlit_option_menu: An additional library for Streamlit that enables the creation of custom dropdown menus with multiple options.pydub: A library for audio file manipulation, supporting various operations such as cutting, merging, and volume adjustment.
The ffmpeg application must be installed on the machine for converting various audio types and is a requirement for using the whisper library. Download it from: https://ffmpeg.org. On MacOS, version 5.2 can be installed via Homebrew with brew install ffmpeg.
To download the Docker image, Docker must be installed. Use the following command in the terminal:
docker pull andrecorumba/leia-docker