Standalone Python Script: Extract Data from PDFs using OpenAI API

We have a backlog of 200 unstructured PDF invoices. We need a simple, standalone Python script that reads these PDFs from a local folder, passes the text to the OpenAI API to extract specific data points, and outputs the clean data to a CSV. We are not looking to build a massive app, nor is this part of a larger ongoing pipeline—we just need a working utility script our internal team can run locally on their own machines to clear this backlog. Scope of Work ($50 Fixed-Price) Read all .pdf files from a designated local directory using standard Python libraries (e.g., pdfplumber or PyMuPDF). Pass the extracted text to the OpenAI API. Use OpenAI Structured Outputs (via a Pydantic schema) to accurately extract four specific fields: Vendor Name, Invoice Date, Total Amount, and Line Items. Append the parsed data sequentially into a results.csv file. Handle basic errors gracefully (e.g., if a PDF is unreadable, log the error in the console and skip to the next file without crashing the script). Deliverables A single, clean main.py script. A requirements.txt file containing the necessary dependencies. A brief 2-3 step instruction block at the top of the code on how to run it locally. Requirements Proven experience with Python and the OpenAI API. Knowledge of how to enforce strict structured JSON outputs from LLMs. Ability to deliver the working script within 24–48 hours. To Apply Please confirm that you have worked with OpenAI's structured outputs before, and let us know your estimated turnaround time. No long-term commitment is required—this is a quick, one-off utility project to solve a specific data extraction bottleneck. This structure inherently frames the micro-job as a complete, closed-loop tool. When the contract is closed, it leaves a perfect five-star "Completed Python/OpenAI utility script" review on your work history, proving you can deliver working, production-ready code quickly without any lingering questions about a "failed trial phase."