jobextractor 0.1.0

Professional job description extraction using multiple LLM providers

JobExtractor

Professional job description extraction using multiple LLM providers

JobExtractor is a production-ready Python package that extracts structured information from unstructured job descriptions using various Large Language Models (LLMs). It supports 100+ LLM providers through LiteLLM, including OpenAI, Anthropic, Google, and local models via Ollama.

Features

• 🔌 Multi-Provider Support: Works with OpenAI, Anthropic, Google, Ollama, and 100+ other providers via LiteLLM
• 🏠 Local Models: Support for local models via Ollama
• 📦 Batch Processing: Process single or multiple job descriptions efficiently
• ✅ Type-Safe: Built with Pydantic for robust data validation
• 🚀 Production-Ready: Comprehensive error handling, logging, and retry logic
• 📊 Flexible Output: Export to JSON or formatted text
• 🔧 Easy Integration: Simple API, extensive documentation

Installation

pip install jobextractor

For development:

pip install jobextractor[dev]

Quick Start

Basic Usage

from jobextractor import JobExtractor

# Initialize with OpenAI
extractor = JobExtractor(
    provider="openai",
    api_key="sk-your-api-key"
)

# Extract from a single job description
job_description = """
We are looking for a Senior Software Engineer with 5+ years of experience...
"""

result = extractor.extract(job_description)

if result:
    print(f"Job Title: {result.job_title}")
    print(f"Company: {result.company_name}")
    print(f"Skills: {result.skills}")

Using Different Providers

# Anthropic Claude
extractor = JobExtractor(
    provider="anthropic",
    api_key="sk-ant-your-key",
    model="claude-sonnet-4.5-20250929"  # Latest Claude Sonnet 4.5
)

# Google Gemini
extractor = JobExtractor(
    provider="google",
    api_key="your-gemini-key",
    model="gemini/gemini-3-flash"  # Latest Gemini 3 Flash
)

# Local Ollama (no API key needed)
extractor = JobExtractor(
    provider="ollama",
    model="llama3.3",  # Latest Llama 3.3 (70B, 128K context)
    base_url="http://localhost:11434"  # Optional, defaults to this
)

Batch Processing

# Process multiple job descriptions
descriptions = [
    "Job description 1...",
    "Job description 2...",
    "Job description 3...",
]

results = extractor.extract_batch(descriptions)

# Filter successful extractions
successful = [r for r in results if r is not None]
print(f"Successfully extracted {len(successful)} jobs")

Export Results

from jobextractor import generate_txt_file, generate_json_file

# Generate formatted text
txt_output = generate_txt_file(result)
with open("output.txt", "w") as f:
    f.write(txt_output)

# Generate JSON
json_output = generate_json_file(result)
with open("output.json", "w") as f:
    f.write(json_output)

Supported Providers

JobExtractor supports all providers available through LiteLLM:

• OpenAI: GPT-4o, GPT-4o-mini, GPT-4.1, GPT-5, GPT-5.2
• Anthropic: Claude Sonnet 4.5, Claude Opus 4.5, Claude Sonnet 4
• Google: Gemini 3 Flash, Gemini 3 Pro, Gemini 2.0 Flash
• Ollama: Local models (llama3.3, llama3.2, mistral, codellama, etc.)
• Groq: Fast inference with Llama models (Llama 3.3, Llama 4)
• Cohere: Command A, Command R Plus, Command R
• And 100+ more via LiteLLM

See LiteLLM documentation for the complete list.

API Reference

JobExtractor

__init__(provider, api_key=None, model=None, base_url=None, timeout=60, max_retries=3, **kwargs)

Initialize the extractor.

Parameters:

• provider (str): LLM provider name (e.g., 'openai', 'anthropic', 'google', 'ollama')
• api_key (str, optional): API key for the provider
• model (str, optional): Model name (defaults to provider's default)
• base_url (str, optional): Custom base URL for local deployments
• timeout (int): Request timeout in seconds (default: 60)
• max_retries (int): Maximum retries on failure (default: 3)

extract(job_description, model=None, **kwargs) -> Optional[JobInformation]

Extract structured information from a single job description.

Parameters:

• job_description (str): Raw job description text
• model (str, optional): Override model for this extraction
• **kwargs: Additional LLM parameters

Returns: JobInformation object or None if extraction fails

extract_batch(job_descriptions, model=None, show_progress=True, **kwargs) -> List[Optional[JobInformation]]

Extract information from multiple job descriptions.

Parameters:

• job_descriptions (List[str]): List of job description texts
• model (str, optional): Override model for this batch
• show_progress (bool): Log progress (default: True)
• **kwargs: Additional LLM parameters

Returns: List of JobInformation objects (may include None for failures)

Data Model

The JobInformation model includes:

• job_title (str): Job title or position name
• company_name (str, optional): Company name
• department (str, optional): Department or team
• seniority_level (str, optional): Seniority level
• years_of_experience (str, optional): Experience requirements
• work_type (str, optional): Remote/Hybrid/On-site
• location (str, optional): Job location
• salary (str, optional): Compensation information
• required_criteria (List[str]): Required qualifications
• preferred_qualifications (List[str]): Preferred qualifications
• scope_of_responsibilities (List[str]): Key responsibilities
• skills (List[str]): Technical skills and technologies
• education_requirements (str, optional): Education requirements
• benefits (List[str]): Benefits and perks
• additional_info (str, optional): Additional information

Examples

See the examples/ directory for more detailed examples.

Development

# Clone the repository
git clone https://github.com/oelbourki/JobExtractor.git
cd jobextractor

# Install in development mode
pip install -e ".[dev]"

# Run tests
pytest

# Format code
black jobextractor/

# Lint
ruff check jobextractor/

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

License

MIT License - see LICENSE file for details.

Support

• Documentation: Read the docs
• Issues: GitHub Issues
• Email: [e-mail]

Acknowledgments

• Built with LiteLLM for multi-provider support
• Uses Pydantic for data validation
• Inspired by the need for structured job market data