v1.0.1

Fix bug on how generated sequences are trimmed among generative steps.

v1.0.0

This new version incorporates the following changes:

New model weights

The model has been retrained on a 5x larger dataset of subdomains and now uses the apex domain, including the TLD, for prediction.

Recursive resolution

Introduces a recursive loop where subdomains found at iteration k are automatically fed as input to iteration k+1.
This process continues until no new subdomains are discovered or a maximum recursion depth is reached (input parameter max-recursion), allowing you to uncover a much deeper and more comprehensive list of subdomains.

Multi-apex support

The model now allows for running inference with multiple apex domains. Use flag --multi-apex.

Miscellaneous quality of life improvements

Silent (-s) and quiet (-q) flags for less verbose outputs.

v0.5.1

We added support for Python3.9 by adding from __future__ import annotations to subwiz/types.py.

We also added tests that check that subwiz runs in all officially supported versions of Python.

v0.5.0

Added two new arguments:

--quiet or -q which only prints the found domains and no pretty logs (useful for piping into another tool).
--silent or -s which outputs nothing to stdout. (must be used together with the -o or --output-file args, so that results are stored in a file rather than in stdout).

⚠️ Warning:
The previous -q flag (max_new_tokens) has been changed to -m.

v0.4.1

This release includes four updates:

• Aligned outputs: The formatting of outputs has been correctly tabbed.
• Bug fix and new tests: #21
• Input no subdomains: You can now provide inputs without subdomains and the model will by itself (probably) try the most likely subdomains.

Execution result from hadrian.io as input:

v0.4.0

🔃 This new version introduces a recursive loop where subdomains found at iteration k are automatically fed as input to iteration k+1. This process continues until no new subdomains are discovered or a maximum recursion depth is reached (input parameter max-recursion), allowing you to uncover a much deeper and more comprehensive list of subdomains.

💣 Example execution: starting only from the www subdomain of a renown social media page we are able to find 1200 different subdomains.

v0.3.0

the new --multi-apex flag allows for running inference with multiple apex domains

v0.2.0

0.2.0

v0.1.5

define python<3.13 since torch does not support 3.13

v0.1.4

Changed type hints to be compatible with Python 3.9.