PDF tools
Compress · Merge · Split · Edit · Sign · Convert
All PDF processing uses pdf.js and pdf-lib running in your browser. Your PDF is parsed and re-serialized locally: the bytes never leave your tab.
Verifiable Privacy Proof
How We Protect Your Files
Most web converters ask you to trust their privacy policies. We prefer to show you verifiable proof. The live network monitor below captures your browser Network activity in real time, showing every request made by this page and confirming zero file uploads.
Live network activity
Live network activity on this page
Files uploaded
0
never
Bytes sent of your data
0 B
since page load
Simulated tool actions
0
local only
Waiting for activity…
Try this yourself: open your browser's Network tab, filter by Fetch/XHR, then use any TryDocsy tool. You'll see static assets load once, with zero requests carrying your file.
What happens when you drop a file
A step-by-step breakdown of our local execution pipeline from file selection to local download.
01
Read into tab memory
When you select or drag a file into a tool, your browser loads the bytes into tab memory using the HTML5 File API and FileReader. The file data remains inside your browser tab's allocated sandbox memory. No network connection is opened.
02
Processed by client-side code
Client-side JavaScript and WebAssembly transform the data directly on your device. PDF documents are parsed and rebuilt with pdf-lib and pdf.js. Images are modified on an HTML5 Canvas. OCR runs through Tesseract.js in WebAssembly, and background removal runs the IS-Net model locally via ONNX Runtime Web.
03
Download created locally
Once processing completes, the output bytes are converted to an in-memory Blob object. The download link calls URL.createObjectURL(blob), prompting your browser to save the file straight from RAM to your disk without any server transfer.
04
Memory freed on reset
When you click reset, load a different file, or close the browser tab, the browser automatically revokes the temporary Blob URL via URL.revokeObjectURL and clears the array buffers through standard garbage collection.
The proof checklist
Six claims, each with a 30-second verification step you can test yourself. The live demo above confirms the first item in real time.
01Your file is never uploaded.
Open DevTools, select the Network tab, and filter by Fetch/XHR. Run any tool with a real file. You will see zero requests transmitting your file contents.
02Tools you have used keep working without internet.
Use a tool once while online, then disconnect from Wi-Fi or turn on airplane mode. The tool still opens, converts, and downloads your file. OCR requires its language model from that first online run; background removal needs an active connection to load its neural network weights.
03No analytics on file contents or document names.
Google Analytics records page views and anonymous tool events: the tool name, the file extension, and a coarse size bucket. It never receives your file names or document contents. You can block analytics using any standard content blocker.
04Source code is auditable in your browser.
Open DevTools and go to the Sources tab. The bundled JavaScript is readable and searchable. The processing logic for each tool lives in the client bundle and runs locally.
05No account, no sign-in, and no email required.
There is no signup form or login modal. Nothing ties your tool usage to an identity, so there is no personal profile to compromise.
06Downloads happen directly on your device.
The download button uses a local Blob URL. Your output file is assembled in browser memory and passed directly to your browser save dialog without leaving the tab.
Independent Verification
How to check it yourself
You do not have to take our word for it. Here are two independent methods to verify that your files never leave your device.
1Monitor the DevTools Network Tab
Open Developer Tools by pressing F12 (or right-click and choose Inspect). Click on the Network tab and select the "Fetch/XHR" filter. Drop any document into a tool and execute the conversion. Notice that no request contains your file data, and zero upload payload bytes leave your browser.
2Test in Offline Airplane Mode
Open a tool you have previously used while online. Then turn off your Wi-Fi or enable airplane mode. Drag in your document and click process. The tool will parse, convert, and download your finished file completely offline. A server-reliant converter cannot function without an internet connection.
Where the privacy comes from
Each tool category runs on an in-browser engine. Here is how client-side technology replaces remote servers.
Image tools
Compress · Resize · Crop · Remove BG · Metadata Remover
Image processing uses the Canvas API and on-device AI models. Your photo is decoded, resized, and re-encoded in browser memory without sending pixels to a server.
OCR tools
Image to Text · PDF to Text
OCR executes via Tesseract.js, a WebAssembly compilation of Tesseract, on your processor. The language model is fetched once and runs completely client-side.
Form tools
Signature Resizer · Passport Photo · Exam Form Resizer
Form utilities process dimension adjustments and binary search kilobyte compression locally. Resized photos and signatures are generated as local Blob downloads.
Network Disclosures
Network exceptions stated honestly
While user documents are never uploaded, TryDocsy makes external network requests for static assets, machine learning weights, and communications. Here is the complete list of outbound requests:
OCR Language Data
Inbound model download only/api/tessdata/[tier]/[lang] & /api/ocr-models/...
When running text recognition, your browser fetches trained language model files from our server. These are static model weights coming down to your device. Your uploaded document is processed by Tesseract.js in your browser and never leaves your computer.
Background Removal Model
Inbound neural network weightsstaticimgly.com
The background removal utility loads the IS-Net segmentation model from staticimgly.com into browser storage on first use. All image processing calculations occur locally on your machine using ONNX Runtime Web.
Contact Form Submissions
Outbound user message/api/contact (delivered via Resend)
Submitting an inquiry through our contact form sends your name, email address, and message text to /api/contact, which transmits it to our inbox using Resend. Messages leave your browser by design.
Product Hunt Badge
Inbound badge imageapi.producthunt.com
The home page loads an external badge image from api.producthunt.com to display community feedback and ratings.
Analytics and Advertisements
Outbound telemetry & inbound adsGoogle Analytics (GA4) & Google AdSense
Google Analytics tracks page views and anonymous tool events (the tool name, the file extension, and a coarse file size category). Google AdSense serves display ads to support free hosting. Neither service ever reads or receives your file names or document contents.
Frequently asked questions about file security
Aren't page assets network requests too?
Yes. When you open any webpage, your browser fetches HTML markup, CSS styling, JavaScript code bundles, and fonts. These requests download application code to your browser. The critical difference is direction: application code comes down from the web server, whereas your personal files and document bytes are never sent back up.
Can I really use TryDocsy offline?
Yes, for tools you have used while online. Our service worker caches the necessary scripts and pages in your browser. If you turn on airplane mode or disconnect from Wi-Fi, those cached tools continue to parse, edit, and convert documents. OCR requires an initial download of language data, and background removal fetches its model weights from staticimgly.com on first use. Running offline proves that processing takes place on your device.
Does TryDocsy use my files for analytics or AI training?
No. Your files never leave your device, meaning they cannot be accessed, stored, or used for model training. Google Analytics records page visits and anonymous tool event tags: the tool name, file extension, and a coarse file size category. It never receives file names or file contents. There is no account system to link usage to an identity.
What if a future tool needs server-side processing?
We adhere to a strict principle: no tool on TryDocsy will upload your file by default. If an advanced feature cannot run in client-side WebAssembly, it will be clearly labeled as an optional server-side utility and documented here so you can decide whether to use it.
How do I audit the client-side code myself?
Open your browser Developer Tools (F12) and inspect the Sources tab. The JavaScript code is served with source maps, allowing you to examine the file manipulation methods in our processing libraries. You can confirm that file operations run directly on ArrayBuffers without making upload network requests.
Ready to test it?
Select any tool, open your browser Network tab, and observe that your file never leaves your machine.
For complete policy details, review our Privacy Policy, our Trust and Security Overview, and our Editorial Policy.