Wikipedia: "Stylometry is the application of the study of linguistic style, usually to written language. It has also been applied successfully to music, paintings, and chess. Stylometry is often used to attribute authorship to anonymous or disputed documents."

That last sentence is important, especially if you are an active or targeted darknet user who comments and/or posts frequently on Dread or other places. Word order and sentence arrangement may act as distinct identifiers of a person.
It should be common knowledge that law enforcement agencies monitor and record every post made here and other forums, even those on the clearnet. Given present AI abilities, Feeding users posts and comments into an AI system to determine unique identifiers is not only possible, it would shock me if they were not doing it already.


Great question. The answer is simple. Run every comment and post you create through a local AI model with a custom prompt before posting anywhere. I'll guide you through the process of setting up your own, step-by-step.

First step, download Ollama.

curl -fsSL https://ollama.com/install.sh | sh

Ollama is an open-source "app" that lets you run, create, and share large language models locally with a command-line interface on MacOS and Linux. All while keeping your chat data away from 3rd party services. Once installed, we are going to pull a model. For this guide, I opt for the neural-chat model, though there are many models selections you can choose from. You can find a full list of supported models on the ollama website. The download is rather large, and may take a while.

ollama pull neural-chat

Now, open your favorite text-editor. Create a new file. For this guide I'll name it 'modelfile'. This will be your model file. Inside the file you can create your custom prompt. Here is an example.

FROM neural-chat

# set the temperature to 1 [higher is more creative, lower is more coherent]
PARAMETER temperature 1

# set the system message
SYSTEM """
You are an anti-stylometry assistant. You will take sentences I give you, and rephrase them to help combat stylometry.
"""

Now save that file. In your terminal, cd into the directory where the model file you just created is located. Run the following command:

ollama create modelname -f ./modelfile

You can change 'modelname' to whatever you wish. That's it. You've just created a local anti-stylometry ai model. Now, let's give it a go:

ollama run modelname

All that's left to do is provide a statement and receive it back, entirely reworded, all within your terminal.


Though not flawless, it performs excellently in concealing sentence structure. With large pieces of text it can struggle. The speed depends on your CPU and GPU (if you have one).

Thank you for reading. You can never be too careful here.