Encode and decode messages using the OBFUS-69 and OBFUS-6969 obfuscation schemes, with AI-assisted decoding powered by your local LLaMA model via Ollama.
- Encode plaintext into OBFUS-69 or OBFUS-6969 obfuscated messages
- Decode obfuscated messages automatically detecting the mode
- Generate all possible decodings and let your local LLaMA AI pick the most likely meaning
- Configurable to select LLaMA model automatically or manually
- GPU saver modes to skip AI calls when unnecessary
- Fully offline decoding using locally hosted Ollama API
-
Install Ollama
Download and install Ollama from https://ollama.ai. Ollama is required to run the local LLaMA models for AI decoding.
-
Download LLaMA model(s)
After installing Ollama, open a terminal and run:
ollama pull llama3.2
You can choose any model from
llama1tollama4depending on your disk space and performance needs. -
Run the script
Clone or download this repository and run:
python obfus69_decoder.py
- Choose to encode or decode messages.
- When decoding, the script auto-detects whether the message is OBFUS-69 or OBFUS-6969.
- The script generates all possible plaintext combinations and uses LLaMA to pick the best guess.
- Access settings to select the LLaMA model, and enable GPU saver modes to optimize performance.
Settings are saved in ~/.obfus69_config.json in your home directory. You can set:
- LLaMA model: Automatically pick the latest model or specify one manually.
- GPU Saver: Skip AI calls if only one decoding exists.
- GPU Saver+: Always skip AI calls and pick the first decoding.
© 2025 MrCookie & ChatGPT
This project is licensed under the GNU General Public License v3.0 (GPLv3).
See LICENSE file for details.
This tool is intended for fun and educational purposes. Use responsibly.