Voice quality diagnostics that end the argument.
Independent voice diagnostics. Defensible evidence.
When calls sound bad, your ISP blames your VoIP provider. Your VoIP provider blames your ISP.
VoIP Test gives you the data to make the problem undeniable.
Your speed test passed. Your calls still sound terrible.
Most online “VoIP tests” are something else in disguise: a rebranded speed test, a loop of timed web requests, or a firewall port check. None of them send voice. They measure how fast your connection can shovel data, and voice quality depends on things they never check.
Wrong protocol entirely
Speed tests use TCP, which retransmits lost packets. VoIP uses UDP, where lost packets are just gone. A speed test literally cannot experience packet loss the way a phone call does.
Jitter without context
Some speed tests report a jitter number, but it is measured against bulk TCP traffic, not the precise 20ms packet cadence that voice requires. A jitter figure with no codec timing, no packet loss correlation, and no MOS impact score tells you nothing about how a call will actually sound.
Buffer bloat hides in plain sight
Speed tests actively trigger buffer bloat, then report the throughput as a success. The 300ms latency spike that would destroy a live call is not on the scorecard.
No picture of real impact
A speed test gives you numbers in isolation. It cannot show you what 2% packet loss at 40ms jitter actually sounds like on a call. Without that connection between metrics and experience, the results do not translate into action.
What’s actually behind a “VoIP test”
Open your browser’s network tab during any free test. What flows during those sixty seconds is what got tested.
| Speed test | Request timer | Port check | VoIP Test | |
|---|---|---|---|---|
| Traffic it sends | Bulk TCP downloads | Loops of timed web requests | A few connectivity probes | Real codec-encoded calls, paced like voice |
| What “jitter” means | Bulk-transfer variance, if shown at all | Request-loop variance, browser overhead included | Not measured | Arrival variation of paced voice packets |
| Packet loss resolution | Hidden: TCP resends lost packets | A handful of probes | Not measured | Counted across thousands of packets |
| Voice under load | No | No | No | Up to 4 genuinely concurrent calls |
| The score | Throughput | A proprietary “readiness” number | Pass / fail signals | MOS via ITU-T G.107, plus the raw data |
| The price of “free” | Ads | Often your email address | Often your email address | Nothing: no account, no email |
The long version, with the questions that reveal which kind you’re looking at: Most free VoIP tests don’t test VoIP.
There is more to the story. Read the full breakdown of what speed tests miss and why it matters.
Diagnostics built specifically for voice.
Standards-based scoring
MOS scores calculated using the ITU-T G.107 E-model, the same framework telecom carriers use. Not a proprietary formula. Not a made-up number.
Test under real load
Run up to four genuinely concurrent calls. Each one is its own peer connection, ICE negotiation, and RTP stream, so you measure the contention between them rather than one call's numbers multiplied by four.
Evidence, not opinions
Reports include methodology documentation and raw data. Cite standards. Attach to support tickets. Use in vendor disputes. Defensible by design.
No installs. No microphone. No live call.
Runs entirely in your browser. Nothing to download, nothing to configure, no permissions to grant. Open the page. Run the test. Get your results.
What actually happens when you run it.
A test you can’t inspect is just an opinion with a progress bar. Here is the whole method, in the open.
Your browser generates the audio
The test creates its own reference audio. It never asks for microphone permission, never listens, never records. Nothing about you is in the signal.
It places real calls
The audio is encoded with a voice codec and sent as a real media stream: small packets at a steady voice cadence, exactly the shape of a phone call. Up to four calls at once, each its own independent stream, so busy-office contention is measured rather than guessed.
Every packet is measured
Arrival time, ordering, gaps. Recorded per packet across the whole run. Loss and jitter are computed from thousands of measured packets, not a sample of ten.
The score follows a published standard
Quality is scored with the ITU-T G.107 E-model, the same framework carriers use, and the report keeps the raw measurements alongside it, in a link you can hand to anyone.
Only an extended test truly qualifies a network.
A short test that finds a problem is proof. A clean one proves the network behaved for one minute of one afternoon. The problems that ambush deployments keep schedules (afternoon congestion, backup windows, load that only exists when the office is full) and no snapshot, ours included, can see a pattern it wasn’t running during.
A day in the life of a network that “tested fine.” The clean result is real, and so is everything it wasn’t there to see.
In alpha · launching Q4 2026
The VoIP Test appliance
Plug it into the network you’re qualifying and walk away. It runs the same real-media measurement as the free test, continuously: a 24–48 hour qualification pass before an install, or permanent monitoring with alerting after one. Quality over time, aligned to the clock, every excursion timestamped: the pattern, not the snapshot. Install is one cable; there is no step two.
Built for MSPs and providers who are done arguing with snapshots: “loss spikes every weekday between 2 and 4 PM” is a finding a carrier can act on. “It tested fine when we visited” is not.
Alpha is running now. Leave your email to get in line, or to hear the moment it launches.
More than a diagnostic tool.
VoIP troubleshooting has always been fragmented, tribal knowledge locked inside vendor forums and expensive consultants. We are building the open resource the industry should have had years ago.
133 articles, 7 in-depth series
VoIP fundamentals, troubleshooting methodology, network design, and MSP operations. Written to be useful whether you are deploying your first phone system or running a managed services provider. Explore the knowledge base
Free browser-based tools
Carrier lookup, bandwidth calculator, SIP response code reference, call flow designer, and more. The kind of tools that used to cost money or require a login. See all tools
Real voice diagnostics
The diagnostic engine sends real voice-profile traffic across your connection, measures what happens to every packet, and produces standards-based reports anyone can understand and defend. Try the test
Built for the people stuck in the middle.
Businesses
Stop saying “my calls sound bad.”
Start sending proof. A diagnostic report with real numbers, the loss pattern behind them, and a shareable link that makes the issue impossible to dismiss.
MSPs & IT Providers
Pre-qualify networks. Win disputes with data.
Assess connections before VoIP rollouts. Deliver professional diagnostic reports to clients. When a vendor points the finger, respond with evidence that cites ITU-T and RFC, not opinions.
VoIP Providers
Send a link. Get a report back.
Customers run the test themselves. You get a shareable diagnostic report. Reduce support time, document issues objectively, and protect your platform’s reputation.
This is the beginning.
VoIP Test is a platform, not a single tool. We are building the diagnostic and educational layer the voice industry has been missing.
Live now
133 articles across 7 in-depth series covering VoIP fundamentals, troubleshooting, network design, and MSP operations. Free planning tools including carrier lookup, bandwidth calculator, SIP code reference, and call flow designer.
Live now
Browser-based voice quality diagnostics. Standards-based scoring, concurrent call testing, shareable reports. Run the test -- free, no account required. A free account adds saved test history.
In alpha – launching Q4 2026
The testing appliance, detailed above: continuous 24–48 hour qualification runs and permanent monitoring from inside the network, with the option to contribute anonymized metrics that improve diagnostics for everyone. Plus professional reporting and API access for teams who need diagnostics at scale.
On the horizon
A contributor node program. Run our lightweight node at your location and get paid to help us gather public routing and peering data that improves diagnostics globally. No internal network data is ever tracked or shared. Plus extended diagnostics and capabilities beyond what the browser can do.
Common questions about testing call quality
Why do my calls sound bad when my internet is fast?
Because speed and call quality are not the same thing. A speed test measures how much data your connection can move in bulk. Phone calls need something different: small pieces of audio arriving steadily, in order, without gaps. A very fast connection that delivers those pieces unevenly will produce choppy, robotic calls while still passing every speed test you run.
Why is this better than a speed test?
A speed test is not bad at its job, it just answers a different question. It checks that data got through and how much of it fits, which is all a download needs: a download does not care whether a piece took an extra half second, only that it turned up. A call does. Every small piece of audio has to arrive on schedule, and one that turns up late is no more use than one that never came at all: you hear the same gap either way. Some speed tests do show a jitter number, but a number on its own does not tell you whether calls will sound fine or terrible. This test places real calls, scores how they actually sounded, and shows you which problem pulled the score down.
How is this different from other online VoIP tests?
Most pages offering a free VoIP test are one of three things in disguise: a speed test, a loop of timed web requests, or a firewall port check. None of them send any voice traffic, so their numbers describe the wrong thing. This test places real calls (audio encoded with a voice codec, sent as a steady stream of small packets the way a phone call sends them) and measures what happens to every packet. The score comes from the same published standard telecom carriers use, so the result means the same thing to your provider as it does to you.
What does a VoIP test actually measure?
This one places real test calls over your connection and listens to what happens to them. It measures how much audio goes missing, how unevenly it arrives, and how long it takes to get there, then turns that into a single quality score for the call. That is the difference between testing your calls and testing your download speed.
Do I need to install anything or turn on my microphone?
No to both. The test runs entirely in your browser tab, and it generates its own audio, so it never asks for microphone permission and never listens to you or records anything. There is nothing to download, no plugin, no account, and nothing left behind on your machine when you close the tab.
How long does the test take?
A minute for a quick check. You can also run two or three minutes, and up to four calls at the same time, which is the more useful option if calls only go wrong when the office is busy. Longer runs with several calls at once are how you catch problems that come and go rather than problems that are always there.
Can a clean test result prove my network is ready for VoIP?
Not by itself, and it's worth knowing why. A test that finds a problem has proven the problem exists: that evidence is solid. A test that comes back clean has only proven the network was fine during that run. Problems that come and go, like afternoon congestion or a nightly backup hogging the connection, are invisible to any test that wasn't running when they happened. So run the test at the times calls actually sound bad, run it under load with several calls at once, and run it more than once. Truly qualifying a network takes a day or two of continuous measurement, which is what our upcoming testing appliance is built for; it's in alpha now, launching Q4 2026.
Can I send the results to my phone provider?
Yes, and this is usually the point. Every result comes with a shareable link you can paste into a support ticket or an email. It carries the measurements and the time the test ran, which is considerably harder for a provider or an ISP to wave away than telling them your calls sound choppy.
What counts as a good result?
The test gives you a score out of five, and above four means callers will not notice anything wrong. Between 3.6 and four is still fine for most business calls. Below about 3.1 is where people start complaining, and below 2.6 the line is not really usable for work. Whatever your score, the report shows which specific problem pulled it down.
Test your connection now.
The voice quality test is live -- free, no account, no installs. Up to 4 simultaneous calls, up to 3 minutes.
Run the VoIP quality testWant updates as new diagnostics launch? Leave your email.
Know someone fighting their phone provider? Send them the test.