Access the #UNGA81 App.
Visualization of the speeches at the #UNGA81 between 22 - 28 September, 2026.
This is a personal project with the aim to explore the texts in the speeches at the #UNGA81.
The app is a combination of using LLMs to extract information and publicly available datasets. You can take a look at the prompts used here.
Screenshot of the streamlit app:
- Summary of the speech
- Summary in one word
- Risks mentioned in the speech
- List of countries mentioned
- A haiku generated from the speech
- Hashtags generated from the speech
- Fictitional newspaper headlines in different styles
- A message from master Yoda
This repository requires uv and Ollama installed.
curl -fsSL https://ollama.com/install.sh | sh
ollama pull llama3.2
ollama pull qwen2.5:3b
ollama pull qwen3.5:0.8b
ollama pull gemma3:4b# Clone this repository
git clone https://github.com/darenasc/unga81.git
# Change directory to repository
cd unga81
# Install dependencies
uv sync
# Run the streamlit app
uv run streamlit run app/app.py
# Download video URLs, transcripts and country information
# uv run unga81- Ollama
- sqlite3
- Vector data from Natural Earth Data
- Worldometer.info
- Sustainable Development Report database
- #UNGA81 - United Nations General Assembly - YouTube playlist
- UNGA81 Speech urls
- Admin 0 – Countries
- Sustainable Development Report - Access full database in Excel (accessed 27.09.2026)
- iso3166-flags (flags)
Create an issue with your suggestions and recommendations.
This repository contains a project designed to analyze and process speeches from the United Nations using Large Language Models (LLMs). The purpose of this project is for entertainment and educational purposes only.
No Endorsement or Guarantee
The analysis provided by this project should not be considered as an endorsement or guarantee of the accuracy, completeness, or reliability of the information contained within. This project is intended to provide a general overview of speech patterns and trends, but it may not capture the full complexity and nuance of the original speeches.
Limitations and Caveats
- The analysis is based on pre-existing data and models, which may be biased towards certain perspectives or interpretations.
- The LLMs used in this project are trained on vast amounts of text data, including texts that may contain errors, inaccuracies, or outdated information.
- This project should not be used for decision-making or policy purposes, as the analysis may not provide a comprehensive understanding of the complex issues involved.
No Liability
The repository author and maintainer make no warranties, express or implied, regarding the accuracy, completeness, or reliability of the information contained within. By using this project, you acknowledge that you are using it for personal, non-commercial purposes only.
Attribution
By using this project, you agree to hold harmless the repository author and maintainer from any claims or liabilities arising from the use of this project.