https://github.com/sverrenystad/meeting-mate
Meeting Mate is a local tool for transcribing and summarizing meetings conducted in Norwegian.
https://github.com/sverrenystad/meeting-mate
asr local-llm meeting-minutes nbailabbeta ollama summarization transcription
Last synced: 4 months ago
JSON representation
Meeting Mate is a local tool for transcribing and summarizing meetings conducted in Norwegian.
- Host: GitHub
- URL: https://github.com/sverrenystad/meeting-mate
- Owner: SverreNystad
- Created: 2023-11-21T16:51:12.000Z (almost 2 years ago)
- Default Branch: main
- Last Pushed: 2025-03-19T20:35:07.000Z (7 months ago)
- Last Synced: 2025-07-01T03:59:37.805Z (4 months ago)
- Topics: asr, local-llm, meeting-minutes, nbailabbeta, ollama, summarization, transcription
- Language: Python
- Homepage:
- Size: 159 KB
- Stars: 2
- Watchers: 1
- Forks: 0
- Open Issues: 1
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# Meeting Mate
Meetings are an essential part of any organization as some work can only be done in them, but they can be time-consuming and unproductive.
One of the key reasons is that decisions and action items can be forgotten or misinterpreted. Sometimes critical information is lost. Either one member needs to take notes and not participate fully in the meeting, or the meeting needs to be recorded and transcribed later. Both of these options are not ideal.
Meeting Mate is a local tool for transcribing and summarizing meeting audio. It leverages cutting-edge Automatic Speech Recognition (ASR) models from `NbAiLabBeta` – fine-tuned versions of OpenAI's Whisper – alongside the Hugging Face Transformers library for transcription. Additionally, it utilizes Mistral and Ollama to generate concise meeting summaries.
**Key Features:**
- **Local Processing:** Runs entirely on your machine with no need for an internet connection after initial setup.
- **Transcription:** Transcribes meeting audio to text giving you a written record of the meeting as well as the audio file.
- **Multiple Interfaces:** Offers both web and desktop GUI options
- **Efficient Summaries:** Summarizes long meetings
## Pre-requisites
Before getting started, ensure that you have the following installed:
- **Ollama and Mistral:**
Download from [Ollama](https://ollama.com/download).
For Linux, you can install Ollama by running:
```bash
curl -fsSL https://ollama.com/install.sh | sh
```
Then download the Mistral model by running:
```bash
ollama pull mistral
```
This will download a 4GB model.
- Install the required Python packages:
Recommended to create a virtual environment before installing the packages.
```bash
pip install -r requirements.txt
```
## Usage
One can use one of two gui interfaces to interact. Either the web interface or the desktop interface.
To run the web interface, run the following command:
```bash
streamlit run streamlit_gui.py
```
or using docker-compose:
```bash
docker compose up --build
```
Once started, access the web interface by navigating to http://localhost:8501 in your browser.
To run the desktop interface, run the following command:
```bash
python tk_gui.py
```