Is My ISP Throttling Me? How to Test on Windows

Your ISP is likely throttling you if certain types of traffic — like BitTorrent, a VPN, video streaming or game downloads — are consistently much slower than your overall connection speed, while a plain speed test still looks normal. The way to confirm it is to compare a clean baseline transfer against protocol-specific transfers: if some protocols are capped while others are not, that gap is the signature of throttling.

Throttling is hard to spot with a regular speed test because most speed tests use plain HTTPS to a nearby server — exactly the kind of traffic ISPs leave untouched. To catch throttling you have to test the traffic types ISPs actually target.

What ISP throttling actually is

ISP throttling is the deliberate slowing of specific traffic rather than your whole connection. An ISP might cap peer-to-peer protocols to manage congestion, slow video streaming to push a paid add-on, or rate-limit VPN traffic. Because only certain protocols are affected, your headline speed test can still report your full plan speed while real-world activities feel slow.

This is different from general congestion (everything is slow at peak hours) and different from a bandwidth cap (you're slowed only after using a data quota). Throttling targets the protocol or destination, not the total volume.

How to test for ISP throttling

The reliable method is a controlled comparison. First, measure a baseline using a plain encrypted transfer to a known-good server — this represents the speed your ISP is happy to give you. Then run transfers that imitate the protocols ISPs commonly target and see whether any of them fall well below that baseline.

NetVeil automates exactly this. Its ISP Throttle Detection runs real protocol transfers and compares each against your HTTPS baseline:

  • BitTorrent — a real BitTorrent wire-protocol transfer, the classic throttling target.
  • QUIC (UDP) — modern HTTP/3 and game traffic increasingly rides on QUIC.
  • WireGuard (UDP) — to see whether VPN-style traffic is rate-limited.
  • OpenVPN — a second VPN fingerprint, since some ISPs target one and not the other.

Reading the results

If every protocol lands close to your baseline, you're almost certainly not being throttled. If one or more protocols are dramatically slower — while the baseline stays high — that's strong evidence of protocol-specific throttling. NetVeil also runs ISP Interference Detection for related tactics like DNS hijacking, captive portals and TLS/SNI inspection, which sometimes accompany throttling.

Because results are graded and exportable, you can attach a report to an ISP support ticket instead of arguing from a feeling. Everything is measured locally on your PC — there's no account and your results aren't uploaded anywhere.

Test your connection

NetVeil is a free, open-source network diagnostics app for Windows. Download it, run the ISP Throttle Detection tool, and you'll have a clear, graded answer in a few minutes.

Frequently Asked Questions

Can a normal speed test detect ISP throttling?

Usually not. Standard speed tests use plain HTTPS to a nearby server — the traffic ISPs leave alone. To detect throttling you need to test protocol-specific traffic (BitTorrent, VPN, QUIC) and compare it against that baseline.

What traffic do ISPs throttle most often?

BitTorrent and other peer-to-peer protocols are the classic targets, followed by VPN traffic (WireGuard/OpenVPN) and sometimes video streaming. Throttling is protocol- or destination-specific rather than a cap on your whole connection.

Is testing for throttling against the rules?

No. Running diagnostic transfers from your own connection is ordinary network testing. NetVeil measures locally and doesn't bypass anything — it just compares how different protocols perform.

How do I prove throttling to my ISP?

Run a repeatable, graded test that shows specific protocols falling below your HTTPS baseline, and export the report. A documented comparison is far more convincing to ISP support than describing that something 'feels slow'.