Cap upload bandwidth via maxUploadBytesPerSecond

One option, no plugin. Set maxUploadBytesPerSecond and the chunked uploader gates each chunk through a rolling 1-second token bucket so the average outgoing rate stays at or below the cap. Useful for being a good citizen on shared infrastructure, throttling background uploads behind interactive ones, or staying under per-IP limits at the load balancer.

Configuration

new MultipleUpload('#uploader', {
 uploadUrl: '/api/upload',
 chunked: true,
 chunkSize: 256 * 1024,
 maxUploadBytesPerSecond: 256 * 1024 // 256 KB/s cap
});

How it works

• The uploader keeps a rolling 1-second window of recent chunk timestamps + sizes.
• Before dispatching each chunk, it sums the window. If sum + chunkSize > cap, it sleeps just long enough to bring the rate back under (max 2-second sleep per chunk).
• Sleeps respect cancel / pause - aborted tasks don't accumulate dead promises.
• Throttle state is per-uploader, not global - multiple uploaders on the page run independently.

Hooks for custom strategies

Custom strategies registered via registerStrategy can opt in by calling uploader._waitForBandwidth(chunkSize) before dispatching:

MultipleUpload.registerStrategy('my-transport', {
 upload: async function (task, uploader) {
 for (const chunk of slice(task.file)) {
 await uploader._waitForBandwidth(chunk.size);
 await fetch('/my-endpoint', { method: 'POST', body: chunk });
 }
 }
});

Use cases

• Background sync: low cap when the user is interactively working in the same app, lift it when the tab is idle.
• Dev-host friendly: 1 MB/s cap during local development to mimic real-world WAN.
• Mobile-friendly defaults: detect navigator.connection?.saveData and set a low cap for users on metered connections.
• Fairness: when a single user uploads many files in parallel, cap each one so the queue progresses smoothly instead of one giant file dominating.

None of Uppy, FilePond, Dropzone, or FineUploader expose a built-in bandwidth cap. This is unique to MultipleUpload.