The Network-tab test: which file converters upload your files?

What "uploads your file" actually means

There are two ways to build a file converter, and which one you are using decides whether your file leaves your device. A server-based converter sends your file to a remote machine, converts it there, and returns the result. Almost every online converter works this way, and it is a normal, legitimate design. An in-browser converter runs the conversion on your own device in WebAssembly, so the file is never uploaded.

The difference is not about which company is more trustworthy. With a server-based tool you are trusting a stated handling and deletion policy you cannot independently check. With an in-browser tool there is nothing to upload, so there is nothing to intercept, retain, breach, or subpoena, and you can confirm that yourself in about a minute. The useful question for any given file is simply: do I want a promise I trust, or a fact I can verify? This page shows the test that turns it into a fact, starting with the result we captured on our own tool.

We ran the test on our own converter, and here is what we captured

This is the part we can show you with evidence rather than assertion. We converted a 1.98 MB PNG to JPG on hushvert with the browser Network tab open and recorded every request. The result: not one request carried the file's bytes (nothing anywhere near 1.98 MB left the browser), and no third-party host was contacted at all. The conversion happened entirely on the device in WebAssembly.

The reason we can promise this and not just claim it is that a continuous-integration test runs on every single code change and fails the build if any request during an in-browser conversion carries more than about 2 KB of body, or if the page contacts any third-party host during that conversion. So the guarantee cannot quietly regress between releases: if a future change ever made an in-browser conversion phone home or leak a file byte, the build would go red before it could ship. The conversion engine is open source under the MIT license, so you can also read exactly what runs in your tab. This is the honest core of the whole thing: for the conversions a browser can do, the no-upload claim is checkable, not promised.

One honest boundary: a few conversions genuinely cannot run in a browser (office documents to PDF, PDF to Word, HTML to PDF, and large video). hushvert runs those on a clearly labeled server lane that tells you 'this one leaves your device' before you start, deletes inputs immediately after conversion, and deletes outputs within about an hour. Everything a browser can do (images including HEIC, audio, archives, small video, and PDF page operations) stays on your device.

The method, so you can reproduce it on any converter

You do not have to take our captured result on faith, and the same test works on any converter, including the big server-based ones. It takes about a minute.

1. Open the converter's page and open your browser developer tools (press F12, or right-click and choose Inspect; on Mac, Option+Command+I). 2. Click the Network tab and tick 'Preserve log' so requests are not cleared. 3. Select your file and run the conversion. 4. Watch the rows that appear. You are looking for an outbound request (usually a POST or PUT) whose body is roughly the size of your file.

How to read it: a real upload is large and scales with your file. Convert a 2 MB photo and an upload shows a request body near 2 MB; convert a 200 KB file and it is near 200 KB. Page assets and a one-time WebAssembly or codec download are downloads, not uploads, and an analytics ping is a tiny outbound request that does not grow when your file does. If you convert a 2 MB file and nothing close to 2 MB leaves your machine, your file was not uploaded.

The even simpler version needs no developer tools. Load the converter page fully while online, then turn on airplane mode or disconnect from the network and try to convert. If it still works offline, nothing was uploaded, because there was nowhere to send it. If it stalls or errors, it needed a server. hushvert ships a live version of this exact test on its privacy-proof page: it loads the codec, invites you to disconnect, converts a locally drawn image offline, and counts every network request the conversion makes.

We ran the same test on the mainstream converters (July 2026)

We did not stop at our own tool. We put the same 1.98 MB image through the popular server-based converters with the Network tab open and watched what actually crossed the wire. Each one uploaded the file. On CloudConvert, selecting the file and clicking Convert sent it as a 1.98 MB POST to its storage servers (eu-central.storage.cloudconvert.com) before conversion. On Zamzar, clicking Convert sent a 1.98 MB POST to zamzar.com and moved to a server-side progress page. On Convertio, the conversion ran on its servers and returned a server-hosted download link, and its own page states the model plainly: 'Cloud-Based Processing: conversion runs on remote servers.'

This is not an accusation. These are legitimate, established services, and server conversion is the normal, and often the only, way to do a job a browser genuinely cannot. The point is that the upload is real and observable, not a matter of opinion: run the one-minute method above on any of them and you will see the same file-sized request leave your machine. Other well-known converters such as FreeConvert, Smallpdf, and iLovePDF are server-based by the same architecture, and each has its own stated deletion policy worth re-checking on its own site. Architectures can change, so the honest thing is a test you can re-run yourself on any tool, on any day, rather than a table to trust. VERT is one of the few exceptions: an open-source, in-browser converter that passes the same no-upload test for the formats it handles.

Which converter should you use? A fair recommendation

Match the tool to the file. For a conversion a browser can do (an image or HEIC photo, an audio re-encode, an archive repack, a PDF merge or split), use a converter that runs locally so the file never leaves your device, and confirm it with the one-minute test. hushvert covers those unlimited and free with no account, and you can verify the no-upload claim end to end. VERT is a solid open-source alternative for the same reason.

For a job a browser cannot do (a Word document to PDF, PDF back to Word, or heavy batch video), you will be uploading no matter which tool you pick, so the question becomes which operator you trust and what they charge. There the mainstream server-based tools are mature and capable, each with real strengths: CloudConvert for breadth and a strong API, FreeConvert for GPU-backed batch video, Smallpdf for a polished PDF experience, Convertio and Zamzar for very broad format coverage. hushvert handles these on its clearly labeled server lane with fast deletion, without pretending the upload is not happening. The honest rule of thumb: for anything sensitive, prefer a converter you can verify does not upload, and accept an upload only when the format genuinely requires it.

Keep reading

• Which file converters upload your files?
• How to verify a converter does not upload
• See the live no-upload proof
• Best private file converter
• File converter FAQ

Common questions

How can I tell if a file converter uploads my files?

Open your browser's developer tools, go to the Network tab, and convert a small file. If you see a request whose body is roughly the size of your file (often a POST to an /upload or storage URL), it is being uploaded. The simpler test: load the page fully, switch on airplane mode, and try to convert. If it still works offline, nothing was sent. Both tests work on any converter, including hushvert.

Did you actually test this, or is it a claim?

Both. We captured it on our own converter (converting a 1.98 MB image in the browser, the Network tab showed no request carrying the file and no third-party host contacted, locked by a CI test that fails the build if it regresses) and on the mainstream ones. Running the same 1.98 MB file through CloudConvert and Zamzar, each sent a roughly 1.98 MB POST to its own servers before converting; Convertio ran the conversion on its servers and returned a server-hosted download. The exact method is in the article so you can reproduce any of it yourself.

Which online converters do not upload your files?

In-browser (client-side) converters do not upload, because the conversion runs in your browser. hushvert runs image (including HEIC), audio, archive, and PDF page conversions entirely on your device, verifiable in airplane mode or the Network tab and locked by a CI test. VERT is another open-source in-browser option. The mainstream server-based converters (CloudConvert, Convertio, Smallpdf, FreeConvert, Zamzar) upload by design and do not offer a verifiable no-upload mode.

Does hushvert ever upload my files?

Only for conversions a browser genuinely cannot do: office documents to PDF, PDF to Word, HTML to PDF, and large video. Those run on a clearly labeled server lane that says 'this one leaves your device' before you start, over an encrypted connection, with inputs deleted immediately after conversion and outputs within about an hour. Everything a browser can do stays on your device and never uploads.

Why do most converters upload instead of running locally?

Server-side is easier to build and can handle every format on a powerful machine. Running the conversion in the browser requires compiling the engine to WebAssembly, which is more work and is only possible for formats a browser can process. hushvert and VERT do this for the formats they cover; many older or broader services were built server-first and have not changed architecture.