DualMindDiffAI

git clone https://github.com/3MPER0RR/DualMindDiffAI.git

cd DualMindDiff

python3 -m venv venv source venv/bin/activate

pip install requests python-dotenv

Usage e configure

insert a sample.txt in data/

export OLLAMA_API_KEY="api_key"

export OLLAMA_BASE_URL="https://ollama.com"

export GOOGLE_API_KEY="api_key"

## Edit model e prompt in main.py



run(
    prompt="Analyze the content in the document.",
    file_path="data/sample.txt",
    safe_llm_mode="ollama-cloud",
    safe_model="gemma4:31b-cloud",
    raw_llm_mode="google",
    raw_model="gemini-2.5-flash"
)

python3 main.py

output

SAFE [filtered analysis from the first model]

RAW [raw analysis from the second model]

DIFF [reasoning differences between the two models]

es CVE-2024-21413

cve-test.mp4