Metadata-Version: 2.4
Name: mask-llm
Version: 0.1.0
Summary: Privacy-safe LLM data cleaning for fine-tuning datasets
Home-page: https://github.com/你的用户名/mask-llm
Author: 你的名字
Classifier: Programming Language :: Python :: 3
Classifier: License :: OSI Approved :: MIT License
Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
Requires-Python: >=3.9
Description-Content-Type: text/markdown
Requires-Dist: spacy>=3.7.0
Requires-Dist: aiohttp>=3.9.0
Requires-Dist: python-dotenv>=1.0.0
Dynamic: author
Dynamic: classifier
Dynamic: description
Dynamic: description-content-type
Dynamic: home-page
Dynamic: requires-dist
Dynamic: requires-python
Dynamic: summary

# mask-llm

Privacy-safe data cleaning for LLM fine-tuning datasets.

Sensitive data (names, emails, phone numbers) is masked **locally** before 
leaving your machine. Only anonymized text is sent to the cloud for cleaning.

## Install

pip install mask-llm
python -m spacy download en_core_web_sm

## Usage

python -m mask_llm \
  --input your_data.jsonl \
  --output cleaned_data.jsonl \
  --key sk-your-api-key

## Output fine-tuning format

python -m mask_llm \
  --input your_data.jsonl \
  --output cleaned_data.jsonl \
  --key sk-your-api-key \
  --finetune

## How it works

1. PII detected and masked locally (never leaves your machine)
2. Anonymized text sent to cleaning API
3. Results restored with original values locally
4. Output as standard JSONL for fine-tuning
