Resume Refinery

Resume Refinery is a CrewAI-powered tool that helps improve resumes while maintaining factual accuracy. It analyzes resumes for ATS compatibility, suggests improvements, and enforces a strict no-fabrication policy.

Features

• Multi-format Support: Accepts plain text, Markdown, DOCX, and PDF resume formats
• ATS Optimization: Analyzes resumes for ATS compatibility and suggests improvements
• Truth Preservation: Strictly prevents fabrication of experiences, dates, companies, or credentials
• Impact Enhancement: Improves wording and structure for maximum impact
• Detailed Reporting: Generates comprehensive reports with suggestions and findings
• Diff Output: Shows changes with clear before/after markers

Installation

pip install resume-refinery

For CrewAI Functionality

To use the full CrewAI functionality with LLMs, you'll need to set up your OpenAI API key:

export OPENAI_API_KEY="your-api-key-here"

Alternatively, you can configure other LLM providers supported by LangChain. The system defaults to using gpt-3.5-turbo for most tasks and gpt-4 for the truth-guarding task for maximum accuracy.

Quick Start

# Basic usage
refine --in my_resume.txt --outdir ./output

# With job description for targeted improvements
refine --in my_resume.md --outdir ./output --jd job_description.txt

# Using CrewAI pipeline
refine --in my_resume.docx --outdir ./output --crew

# Processing a PDF resume
refine --in my_resume.pdf --outdir ./output --crew

Supported Formats

Resume Refinery supports the following input formats:

• Plain text (.txt)
• Markdown (.md)
• DOCX (.docx)
• PDF (.pdf)

Truthfulness Policy

Resume Refinery enforces a strict no-fabrication policy:

• No invented employers, titles, dates, degrees, certifications, or responsibilities
• All changes must be traceable via markup/diff output
• If unknown facts are needed for stronger bullets, placeholders tagged CONFIRM_REQUIRED are used with explanations

Output Files

The tool generates the following output files:

• marked_up_resume.md: Resume with changes marked using [-removed-] and {+added+} syntax
• rewritten_clean_resume.md: Clean, improved version of the resume
• improvement_report.md: Comprehensive report with analysis and suggestions
• rewritten_clean_resume.docx: (Optional) Clean resume in DOCX format if input was DOCX

Architecture

The pipeline consists of several stages:

1. Parsing: Converts input to normalized format
2. Analysis: Validates content and checks ATS compatibility
3. Rewriting: Suggests improvements to wording and structure
4. Truth Guard: Ensures all changes are factual
5. Output Generation: Creates the final artifacts

Contributing

Contributions are welcome! Please submit a pull request or open an issue to discuss changes.

License

This project is licensed under the MIT License - see the LICENSE file for details.