Extract All PDF Text & Numbers
Budget / SalaryHourly project
TypeFreelance project
LocationRemote
Posted2 hours ago
I have a set of PDF files and I need every piece of readable information pulled out—words, phrases, standalone numbers, line items—nothing left behind. The goal is a clean digital dataset that mirrors exactly what appears on each page so I can run further analysis without manually copying anything.
Scope of work
• Parse each PDF and capture all textual and numeric content, page by page.
• Preserve natural reading order and any obvious section breaks so the data stays meaningful once it’s in a spreadsheet or database.
• Flag characters that fail to convert (rare symbols, corrupted glyphs) so I can quickly spot-check them.
• Deliver the final extraction in either .xlsx or .csv format; a JSON export alongside it would be a bonus.
Tools are up to you—Python with PyPDF2, pdfplumber, or an OCR layer such as Tesseract if needed—so long as the output is complete and accurate. If you build a repeatable script, please include clear setup notes so I can rerun it on future documents.
Acceptance criteria
• 100 % of legible text and numbers present in the output.
• Page references or another simple way to trace each entry back to its source page.
• No arbitrary truncation, skipped pages, or merged lines that reduce readability.
Send a short note describing your extraction approach and any similar projects you’ve handled, and we can get started right away.
Scope of work
• Parse each PDF and capture all textual and numeric content, page by page.
• Preserve natural reading order and any obvious section breaks so the data stays meaningful once it’s in a spreadsheet or database.
• Flag characters that fail to convert (rare symbols, corrupted glyphs) so I can quickly spot-check them.
• Deliver the final extraction in either .xlsx or .csv format; a JSON export alongside it would be a bonus.
Tools are up to you—Python with PyPDF2, pdfplumber, or an OCR layer such as Tesseract if needed—so long as the output is complete and accurate. If you build a repeatable script, please include clear setup notes so I can rerun it on future documents.
Acceptance criteria
• 100 % of legible text and numbers present in the output.
• Page references or another simple way to trace each entry back to its source page.
• No arbitrary truncation, skipped pages, or merged lines that reduce readability.
Send a short note describing your extraction approach and any similar projects you’ve handled, and we can get started right away.
Apply on Freelancer →
Project sourced from Freelancer.com. Applications happen directly on the original platform — we never collect your data.