Karib0u / CTrag

A cyber threat intelligence chatbot that ingested 2200+ reports from vx-underground.

CTrag

CTrag is a cyber threat intelligence tool that uses local large language models (LLMs) and a vector database to answer your questions about cyber threats. It's built on top of Langchain, Ollama, Chroma, and PyPDF.

Credits

The majority of the code for this project was adapted from the work of the first developer of local-LLM-with-RAG. The project was then adapted for the specific cyber threat intelligence use-case and a first vector database was built.

https://github.com/amscotti/local-LLM-with-RAG

Requirements

• Ollama version 0.1.26 or higher.
  • You can download other models with ollama pull [MODEL_NAME] so that the chatbot can use them (default: dolphin-mistral)

Setup

1. Clone this repository to your local machine using Git LFS to ensure you receive the already processed vector database with the 2200 reports. If you haven't installed Git LFS, please follow the instructions here.
2. Create a Python virtual environment by running python3 -m venv env.
3. Activate the virtual environment by running source env/bin/activate on Unix or MacOS, or .\env\Scripts\activate on Windows.
4. Install the required Python packages by running pip install -r requirements.txt.

Running the project

Note: The first time you run the project, it will download the necessary models from Ollama for the LLM and embeddings. This is a one-time setup process and may take some time depending on your internet connection.

1. Ensure your virtual environment is activated.
2. Run the streamlit GUI with python streamlit run app.py

Adding reports to the vector database

1. Put the PDFs in the reports folder
2. Run the tool, it will automatically process the files and add them to the db.

Available commands

Here are the available command line arguments and their default values for running the CTrag Streamlit application:

streamlit run app.py [-m MODEL] [-e EMBEDDING_MODEL] [-p PATH] [--nb-docs NB_DOCS]

• -m MODEL, --model MODEL: The name of the LLM model to use. Default is "dolphin-mistral".
• -e EMBEDDING_MODEL, --embedding_model EMBEDDING_MODEL: The name of the embedding model to use. Default is "nomic-embed-text".
• -p PATH, --path PATH: The path to the directory containing documents to load. Default is "reports".
• --nb-docs NB_DOCS: The number of documents to retrieve from the vector database. Default is 8.

Example usage:

streamlit run app.py -m "dolphin-mistral" -e "nomic-embed-text" -p "/path/to/documents" --nb-docs 10

This command runs the CTrag Streamlit application using the "dolphin-mistral" LLM model, the "nomic-embed-text" embedding model, loads documents from the "/path/to/documents" directory, and retrieves 10 documents from the vector database.

Source used to build the original vector database

• VX-Underground archives

Technologies Used

• Langchain: A Python library for working with Large Language Model
• Ollama: A platform for running Large Language models locally.
• Chroma: A vector database for storing and retrieving embeddings.
• PyPDF: A Python library for reading and manipulating PDF files.

License

This project is licensed under the terms of the MIT license.