https://github.com/mvanzulli/meetingassistant.py
A local deployable version of an AI meeting assitant
https://github.com/mvanzulli/meetingassistant.py
meeting-assistant meeting-minutes python speech-recognition speech-to-text summarization
Last synced: 4 months ago
JSON representation
A local deployable version of an AI meeting assitant
- Host: GitHub
- URL: https://github.com/mvanzulli/meetingassistant.py
- Owner: mvanzulli
- License: mit
- Created: 2024-04-21T13:46:58.000Z (over 1 year ago)
- Default Branch: main
- Last Pushed: 2024-04-21T13:51:36.000Z (over 1 year ago)
- Last Synced: 2024-04-21T16:32:44.444Z (over 1 year ago)
- Topics: meeting-assistant, meeting-minutes, python, speech-recognition, speech-to-text, summarization
- Language: Python
- Homepage:
- Size: 868 KB
- Stars: 0
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# 🎙️ Meeting Assistant
This program provides functionality to record meetings and then generate summaries of the transcribed audio using OpenAI's GPT-3 language model.
## 🚀 Prerequisites
To run this program, you will need:
- `Python 3.7` or higher
- `ffmpeg` installed on your system. If not:
```bash
sudo apt install ffmpeg
```
## 🔧 Installation
1. Clone this repository
2. Create an enviroment and install dependencies
```bash
python3 -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt
```
## 🔑 Required Environment Variables
This program requires an OpenAI API key to function. To set up your environment variables:
1. Create a new file named `.env` in the root of this directory
2. Add the following line to your `.env` file: `OPEN_API_KEY=`
3. Replace `` with your actual OpenAI API key.
## 🛠️ Usage
### UI
Launch the Web-UI:
```bash
./run.sh
```
Launch the GUI:
```bash
python3 gui/gui.py
```
### Record
To record a meeting, run:
```bash
python model/model.py record .mp3
```
This will record the audio and save it to a file with the specified name.
### Summarize
To generate a summary of an audio file, run:
```bash
python3 model/model.py summarize .mp3
```
This will generate a summary of the transcribed audio using OpenAI's GPT-3 language model in the same audio language.
### Translate and Summarize
To generate a summary of an audio file and translate it into , run:
```bash
python3 model/model.py summarize .mp3
```
You can see the language keys in the language_roles.yaml file.
## 📝 License
This project is licensed under the MIT License - see the [LICENSE](LICENSE) file for details.
## 💼 Variables to tune
Feel free to change:
This program has several variables that can be tuned to change its behavior. These variables are declared at the beginning of the program:
- `OS`: Set this to `"linux"` or `"MAC"` depending on your operating system.
- `DEVICE`: Set this to `"cuda:0"` if you have an NVIDIA GPU and want to use it to accelerate processing, or `"cpu"` to use the CPU instead.
- `WHISPER_MODEL`: The name of the pre-trained Whisper model to use for transcribing the audio.
- `ENV_OPENAI_KEY`: The name of the environment variable that contains your OpenAI API key.
- `TEMPERATURE`: The "temperature" parameter to use when generating summaries with GPT-3. Higher values will generate more diverse summaries, while lower values will generate more conservative summaries.
- `GPT_MODEL`: The name of the GPT-3 language model to use for generating summaries.
- `GPT_ENCODER`: The name of the GPT-3 tokenizer to use for encoding text.
- `SIZE_CHUNK`: The size of each "chunk" of text to send to GPT-3 for summarization. Larger chunks will result in fewer requests to the API, but may be slower to process.
- Command prompts and command role in the `language_roles.yaml` file.