Real-time load and performance testing for WebSocket and HTTP/HTTPS applications — with a live 3D globe of virtual users converging on your servers.
Watch traffic flow in. See latency climb. Identify the moment your stack starts to fail. Repeatable, measurable, observable.
Pulsar bundles a load generator, a live visualisation engine, multi-phase test plans, and historical analytics into one self-contained tool.
Every virtual user appears as a dot on land, with a glowing arc connecting it to your server. As load ramps up, you see the world fan in — and disappear when connections drop.
Define warm-up, sustained, spike, and cool-down phases per project. Each phase has its own user target, ramp curve, hold duration, and message interval.
p50, p95, p99 are necessary but not sufficient. Pulsar plots the full latency distribution so you can spot bimodal tails and slow paths the percentiles hide.
Pick any two runs from history. Get a delta view of failures, throughput, and a p95 line overlay. Use it for before/after deploys, infra migrations, or A/B configs.
Persistent WebSocket connections with configurable ping payloads. Or HTTP/HTTPS with custom methods, headers, and request bodies. Same UI, same metrics.
Load generation runs in a separate Node.js worker process via IPC, so the UI server stays responsive even at thousands of virtual users hammering your target.
Pulsar geolocates your target by IP and renders an interactive 3D globe. As virtual users come online, animated arcs converge toward your server marker. Disconnects fade out instantly. It's a load test you can hand to a non-engineer and they'll understand the shape of the problem.
Stop guessing what "good" looks like. Pulsar's recommendation engine combines failure rate and p95 latency into a clear traffic-light verdict. Every run is saved with full timeseries, sample failures, and a configuration snapshot — so any verdict is auditable months later.
One URL, one protocol, optional headers, optional body. Save it. Reuse it across runs and deploys.
Warm-up at 50 users, sustained at 500, spike to 2,000, then back down. Pulsar runs the curve and reports the verdict at every step.
Globe shows the shape of the load. Histogram shows the latency distribution. Verdict box tells you what to do next.
Sign in to the dashboard to create your first project and start a run in under a minute.
Sign in →