Why Does Converting a File Online Ever Require an Upload?

How Server-Side Processing Works

In a server-side file tool, the processing happens on a remote computer:

1Server-side processing flow:
2 
3User's device                    Tool's server
4────────────                    ─────────────
51. User selects file
62. Browser uploads file  ──────► 3. File received, stored
7   (network request)                temporarily
8                                 4. Server processes file
9                                    (PDF merge, conversion, etc.)
105. User downloads result ◄────── 5. Server sends result
11   (another request)
12 
13Time = upload time + processing time + download time
14     = (5MB ÷ upload speed) + server processing + (5MB ÷ download speed)
15 
16On 11 Mbps upload: 5MB PDF takes ~3.6 seconds just to upload
17On 2 Mbps (airport WiFi): same file takes ~20 seconds just to upload

How Client-Side Processing Works

1Client-side processing flow:
2 
3User's device (browser)
4──────────────────────
51. User selects file
62. Browser reads file from disk into memory (fast — no network)
73. WASM or JS library processes file in browser memory
84. Result written to disk via download prompt
9 
10Time = file read time + processing time
11     ≈ milliseconds + 1-5 seconds (depending on file size and operation)
12 
13No upload. No download wait. Works offline. File never leaves the device.

Speed Comparison

When Server-Side Is the Right Choice

Client-side processing has real constraints. Server-side is genuinely better for:

• Very large files (500MB+): Browser memory limits make multi-gigabyte files impractical client-side. Servers have no such constraint.
• GPU-accelerated video encoding: H.264/H.265 hardware encoding requires server-side hardware. FFmpeg.wasm is functional but significantly slower than a server with a dedicated encoder chip.
• Server-side secrets: Signing a document with a company-level private key, generating bank-grade signatures, or stamping PDFs with a certified certificate, these require a server holding the private key.
• Complex AI operations: Background removal using a large neural network, high-quality audio transcription at scale, models over 1GB are impractical to run in a browser tab.

When Client-Side Is the Right Choice

For the majority of everyday file operations, client-side wins on every metric except raw compute for very large files:

• Privacy-sensitive documents: Contracts, medical records, financial data, anything you would not want on a third-party server.
• Poor connectivity: Offline work, slow mobile connections, metered data plans.
• GDPR compliance: Eliminates the need for a Data Processing Agreement with the tool provider.
• Reliability: No server outage can affect a client-side tool. No API rate limits. No paywall blocking access.
• Repeated operations: Processing 100 images one by one, each one benefits from zero upload latency.

Hybrid Approaches

Some tools use a hybrid model: process what they can in the browser, fall back to a server for operations that genuinely require it. The key is transparency, the tool should tell you when a file is leaving your browser, rather than silently uploading everything by default.

For an explanation of how to verify what a specific tool is doing with your files, our client-side processing privacy guide covers the DevTools verification method. For the risks associated with tools that process files server-side without transparency, our file upload risk guide covers what you cannot verify about a remote server.