How Novaxera Works
Novaxera processes your file on its own servers with established open-source engines — PyMuPDF, pdf2docx, LibreOffice, Pillow, Tesseract OCR and rembg — and returns the result as a download. Files travel over HTTPS, the original upload is deleted as soon as processing finishes, and the result is deleted automatically within 60 minutes.
What happens to your file
-
You upload a file
Your browser sends the file to Novaxera over an encrypted HTTPS connection. Each file can be up to 50 MB; larger uploads are rejected with a clear message.
-
The server processes it
The file is handled by the engine for that tool (see the table below) in a temporary working folder on Novaxera's server.
-
The upload is deleted
As soon as processing finishes, successful or not, the original upload is deleted from the server.
-
You download the result
A Download button appears with the result file. Tools that create many files deliver a single ZIP.
-
The result is deleted
A cleanup job deletes every result file automatically within 60 minutes, whether or not it was downloaded.
The engines behind each tool
Novaxera is built on established open-source software. This is what runs when you use each tool.
| Tool | Engine | What it does |
|---|---|---|
| Compress PDF | PyMuPDF + Pillow | Downsizes embedded images larger than 1200 px, re-encodes JPEGs at quality 72, removes unused objects and deflate-compresses the file. |
| PDF to Word | pdf2docx, LibreOffice fallback | Rebuilds paragraphs, tables and images from the PDF's text layer into an editable .docx. |
| Word to PDF | LibreOffice | Lays out .docx and .doc files and exports them as PDF. |
| Merge, Split, Protect, Unlock, Watermark | PyMuPDF | Copies pages between documents, applies AES-256 encryption, removes encryption with your password, or writes a diagonal text watermark. |
| PDF to Images | PyMuPDF | Renders every page as a PNG at 150 DPI. |
| Edit PDF | PDF.js (preview) + PyMuPDF | Shows the pages in your browser, then redacts the original line and writes your new text in its place. |
| Compress Image, Convert Image, Images to PDF | Pillow + PyMuPDF | Re-encodes images (JPEG/WebP quality 75), converts between six formats, or places each image on its own PDF page. |
| Remove Background | rembg with the U²-Net model | A deep-learning segmentation model finds the main subject and makes everything else transparent. |
| Image to Text (OCR) | Tesseract OCR | Recognizes printed text in 10 languages, including Arabic, and returns a UTF-8 .txt file. |
Our principles
No accounts
There is nothing to sign up for, so Novaxera holds no profile, password or file history for you.
Short retention
Uploads are deleted after processing and results within 60 minutes. There is no file storage feature.
Files are not reused
Your files are not shared, sold, read by people or used for anything other than producing your result — including AI training.
Documented behaviour
Each tool page states the engine, settings and limits, so you know exactly what will happen to your file.
Honest limits
Where a tool can't do something — such as reading handwriting or watermarking in Arabic — its page says so.
Questions about how Novaxera works
Does Novaxera process files in my browser or on a server?
Novaxera processes files on its own servers. Your browser uploads the file over an encrypted HTTPS connection, the server runs the tool's engine, and you download the result. The PDF editor also shows a preview of your pages in the browser, but the actual edits are applied on the server.
How long does Novaxera keep my files?
Novaxera deletes your original upload as soon as processing finishes, and deletes the result file automatically within 60 minutes. There is no account, history or file storage feature, so nothing is kept after that. Download your result promptly, because the download link stops working once the file is deleted.
Does Novaxera use my files to train AI?
No. Novaxera does not use uploaded files for AI training or for any purpose other than producing the result you asked for. The Remove Background tool runs a pre-trained open-source model, U²-Net, which does not learn from your images.
Which open-source software does Novaxera use?
Novaxera is built on PyMuPDF for most PDF operations, pdf2docx and LibreOffice for Word conversions, Pillow for images, Tesseract for OCR, rembg with the U²-Net model for background removal, and PDF.js for previewing pages in the PDF editor. The website itself runs on Django.
Why does Novaxera have a 50 MB limit?
The 50 MB per-file limit keeps processing fast and reliable for everyone on a free service. Most documents, scans and photos are well under it. If a PDF is larger, Novaxera's Split PDF tool can divide it into smaller files first.
Last updated · Privacy policy · All tools