Privacy & offline
Built so there's nothing to leak
PDFort has no servers and no accounts. Here is exactly what happens on your phone, and the few things that use the internet.
PDFort has no backend. Every tool runs on your phone's processor.
No sign-up, no email, no login. Open the app and start.
You pick each file with Android's own picker. PDFort can't browse your phone.
Internet use
What does use the internet
| What | When | Your documents sent? |
|---|---|---|
| Ads (free version) | Banners on the home and files screens; an occasional full-screen ad after you finish. AdMob may use your advertising ID. | Never |
| Buying, restoring and checking Pro | Through Google Play; our server confirms the purchase with Google using its purchase token | Never |
| Diagnostics | Only if you turn sharing on: daily error-code counts, app and Android version | Never |
| Translation language packs | Once per language you choose (about 30 MB each) | Never |
| Scanner, and Chinese/Japanese/Korean text recognition | Google Play services may download these modules the first time | Never |
App lock
Turn on App lock in Settings and PDFort asks for your fingerprint, face or phone PIN when you open it, and again after a minute in the background. Android handles the check — PDFort never sees your fingerprint.
Private storage
Files you create live in PDFort's private storage, where other apps can't read them. Use "Save to device" for anything you want to keep outside the app — uninstalling PDFort deletes its files.
The honest part
What to expect from on-device processing
Doing everything on your phone is what keeps your files private. It also comes with a few trade-offs we'd rather tell you about up front.
Big files take longer and use more battery
Your phone is doing the work a server farm would. Files over about 150 MB or 300 pages get a warning first; they may be slow, warm the phone, or fail on phones with little memory. Splitting a huge file first usually helps.
Strong compression turns pages into images
"Light" keeps text sharp and selectable. "Balanced", "Strong" and "Fit a size limit" re-encode pages as images, which shrinks scans dramatically but means text can't be selected afterwards.
Text recognition isn't perfect
It's very good on clear printed text, weaker on handwriting, stylised fonts and dim photos. Arabic and Urdu use a separate engine that is slower than the others. Bengali, Tamil, Telugu, Persian and Russian can be translated but not yet read from images.
Translation needs a one-time download
Google doesn't allow its translation models inside apps, so each language (about 30 MB) is downloaded once. You can choose Wi-Fi only and remove packs any time in Settings → Languages.
Redacted pages lose selectable text
That's what makes the redaction permanent: the page is rebuilt as an image with the boxes burned in.
PDF to Word is a best effort
Paragraphs, headings, bold and italic come through. Complex tables and multi-column layouts become plain paragraphs. "Exact layout" keeps the look but isn't editable text.
The app stays small on purpose
Only the most-used text recognition languages are built in. Translation packs are downloaded only for the languages you pick, and PDFort cleans up its temporary files automatically. Settings → Storage shows what it's using.
Read the full privacy policy.