About

Most home-internet advice is either an ISP's sales page or a forum post that assumes you already know the answer. StabilityPulse is three small tools for the in-between: figuring out what your network is actually doing, what to do about it, and whether the plan you're paying for makes sense.

Three pillars

Each one is a standalone tool — free, no signup, browser-only — and they link into each other where it's useful.

Plan

Right-size your internet plan.

A calculator that maps what your household actually does (calls, streams, downloads, gaming, work-from-home) to the bandwidth and stability you actually need — and prints the overbuy in dollars per year if you're paying for more. Backed by a cornerstone post covering household concurrency math, symmetric-vs-asymmetric fiber, and the hidden hardware costs ISPs don't mention when they sell you 5 Gbps.

Diagnose

Run a real 30-second test.

A synthetic WebRTC stream through a public TURN relay — the same transport Zoom, Teams, and Meet use. Measures jitter, packet loss, and loaded latency, then turns the result into per-app verdicts for 8 workloads grouped into Calls, Gaming, and Streaming. Throughput-aware: if you also run the burst test, a workload that needs both bandwidth and low jitter gets graded against both.

Fix

Match your symptom to the fix.

15 symptom fingerprints (“PC Wi-Fi slow but phone fine,” “ethernet capped at exactly 90 Mbps,” “Teams call quality keeps dropping”) indexed by what the problem looks like to you, not by product category. Each one routes into a long-form guide that walks the diagnostic chain — driver, band, cable, router setting, ISP — with the specific knobs to turn.

What the test actually measures

Speed tests like Ookla and Fast.com report peak burst throughput over TCP. A 500 Mbps result means your link can push 500 megabits during a one-shot transfer. Real-time apps don't stream that way: Zoom, Teams, and Meet send small UDP packets at a steady cadence and need every packet to land within tight timing budgets. What breaks calls is jitter, packet loss, and bufferbloat — none of which appear on a speed test result.

Loaded latency: Round-trip time while the link is busy. Calls feel laggy when this exceeds ~150 ms.
Jitter: Variance in packet inter-arrival time. The single biggest predictor of robotic audio.
Packet loss: Fraction of RTP packets dropped. Anything above 1% is audible.
Bufferbloat: Loaded RTT minus idle RTT — extra latency a saturated buffer is adding. Common on consumer routers.

How the score works

The four metrics roll up into a 0–100 score using a weighted model: jitter 40%, loss 40%, RTT 20%. Jitter and loss damage real-time audio more severely than absolute latency does — a steady 200 ms link sounds fine on a call; a 50 ms link with bursts of 500 ms does not.

Green: 80 or higher — calls should feel native.
Yellow: 50 to 79 — usable, but expect occasional artifacts.
Red: under 50 — calls will degrade noticeably.

Per-app verdicts use each app's own published quality thresholds (Zoom, Teams, Meet — links in the test result), so a yellow overall can still be a green for one app and a red for another. Every raw number is shown alongside the verdict; you can disagree with the weighting and still trust the data underneath.

What we deliberately don’t do

No accounts, no history. Results never leave your device except as the WebRTC stream itself. No tracking pixels, no email capture as a gate.
No upsell. Every Fix guide is full content; the calculator is the calculator. Amazon affiliate links exist where a specific product is the honest answer — disclosed inline, never the only option.
No competing speed test. The optional burst test is a contrast feature so you can see the Mbps-vs-stability gap on your own line. It's not meant to replace Ookla.

Honest limitations

The TURN relay sits in one geographic region (Nuremberg). If you're far from it, baseline latency will be higher than what you'd see calling someone nearby. Multi-region relay is on the roadmap.
Background tabs throttle WebRTC stats. Keep the test tab in the foreground while it runs.
A bad result on Wi-Fi doesn't always mean a bad ISP. Try once on Ethernet to separate router/Wi-Fi problems from carrier problems — the Fix matcher will route you to the right place either way.
The Plan calculator is American-centric on hardware costs and ISP behavior. The bandwidth math holds elsewhere; the dollar figures don't.

Who builds this

Solo project, side-built. The product is intentionally opinionated and intentionally free — nobody's paying for it yet and that's fine. Monetization is the Amazon affiliate links inside the Fix guides; no plan to bolt on a subscription unless multi-region monitoring eventually justifies one.

Right-size my plan Run the test →Match my symptom