v0.1.3

Someone forgot to change the toml file.

v0.1.2

Fixed Pydantic pre-release version that was causing problems with pypi users.

v0.1.1

Slight Typo Fix in the README

v0.1.0 - Initial Release

🚀 v0.1.0 - Initial Release

We are excited to announce the first official release of AXEtract, a high-performance, low-cost framework for structured web data extraction. AXEtract implements the methodology described in our research paper, achieving state-of-the-art results using specialized 0.6B parameter models.

✨ Key Features

• Specialized LoRA Adapters: Optimized adapters for DOM pruning and structured extraction, significantly reducing token costs.
• Smart DOM Pruning: High-speed classification and removal of irrelevant HTML noise before extraction.
• Grounded XPath Resolution (GXR): Automatically maps extracted JSON fields back to their original source XPaths for 100% verification.
• Dual-Engine Support: Built-in support for vLLM (production-grade throughput) and Transformers (local research/CPU).
• FastAPI Server: Ready-to-use API server with axe-server CLI.

📦 Installation

pip install axetract

🚥 Quick Start

from axetract.pipeline import AXEPipeline
from pydantic import BaseModel

class Product(BaseModel):
    name: str
    price: str

pipeline = AXEPipeline.from_config(use_vllm=False)
result = pipeline.extract("https://example.com/product", schema=Product)
print(result.prediction)

📚 Documentation

Check out our full documentation at: https://abdo-mansour.github.io/axetract/

📝 Citation

If you use this work, please cite:

@misc{mansour2026axe,
      title={AXE: Low-Cost Cross-Domain Web Structured Information Extraction}, 
      author={Abdelrahman Mansour and Khaled W. Alshaer and Moataz Elsaban},
      year={2026},
      eprint={2602.01838},
      archivePrefix={arXiv},
      primaryClass={cs.CL},
      url={https://arxiv.org/abs/2602.01838}, 
}

---

Full Changelog: https://github.com/abdo-Mansour/axetract/commits/v0.1.0