Documents and notes decision
Sejda Web
A focused PDF→Word web service is feasible for a capable engineer in ~30 hours using open-source libraries (PyMuPDF, python-docx, Tesseract); Sejda's broader product polish, offline desktop app, Drive integrations, and workflow features would require more work to match.
Visit website↗Open-source builds that already do this
Every project below is open source and already does this job today. Fork one, self-host it, or take the parts you need — the build prompt further down assumes an empty file, and this is the shortcut past that. Licences differ; check the one on each card before you ship.
What a replacement has to do
- Upload PDF → extract/convert pages and text → (optional) OCR → assemble .docx → provide download
What it still won’t have
- Offline desktop application with files never leaving the computer (Sejda Desktop)
- Built-in Google Drive/Dropbox integration and UI polish
- Prebuilt workflows/automation and team/volume billing features
- Polished error-handling, edge-case layout fidelity and proprietary conversion heuristics
What remains hard
- Product polish and ongoing maintenance
First-year cost
Keep paying
Paying is—cheaper in year one.
On cash alone, building overtakes the subscription at 1 seat.
Money you would actually spend
Time you would spend
—
What you would spend
What we assumed
The verdict above measures whether you could build it. This one is only about money.
Runnable build prompt
Build a minimal web service (Python FastAPI backend, Redis for short-lived job state, S3-compatible storage for temporary files, React frontend) that converts PDFs to Word: accept multipart uploads and Google Drive/Dropbox URL inputs; extract text and images using PyMuPDF; run Tesseract OCR on pages that lack text; assemble .docx with python-docx preserving paragraphs and embedding images; provide a download link and auto-delete files after 2 hours. Out of scope: advanced layout recovery, WYSIWYG editor, team billing UI, desktop app. Include retries, input validation, per-file size limits, unit tests for core conversion functions, and logging/metrics.
How we checked
How the score was reached
- Build verdict base78
- An open-source build was found+5
- 4 cited sources+3
- 3/3 assessment runs agreed+4
- Evidence score90
The base comes from the verdict. Everything under it is a check that either happened or did not, and each one is a fact frozen in this record rather than a judgement made at render time — so the same evidence always produces the same number.
How scoring works →Cited sources · 4
Every page the run actually retrieved.
- official productSejda home / product pages
- official pricingSejda pricing / upgrade
- official docsSejda PDF to Word feature page
- open sourcePyMuPDF repository
Integrity checks
What held up, and what did not.
