Fire any endpoint.
Watch it shake.
Pulse takes a single endpoint, fires it N times with a tiny bit of concurrency, and turns every response into a living tile. Fast ones glow green. Slow ones shake and grow a snail. Mistakes turn black. No YAML, no k6 install, no grafana dashboards — just a wizard and an honest verdict.
30 days of Pulse,
on the house.
Already on a TestingForge plan? Pulse is included free for the first 30 days — no extra invoice, no setup, no card. Open the workspace, point it at one of your environments, and start firing.
- ✓ Auto-applied to every paid TestingForge workspace.
- ✓ Full feature set — live tiles, p50/p95/p99, letter grade, run history.
- ✓ Fair-use caps to keep things friendly for shared infra.
- ✓ Cancel any time. Hands off your card on file.
A game-like wizard, not a stack trace.
Pulse is not a load-test platform. It's a fast, honest feel for an endpoint — the way a drummer tests a kit before a show. Three steps from zero to verdict.
Pick a target
Choose an existing brick (we copy its URL, method, headers, auth) or just type a URL. Pulse only lets you hit hosts your organization owns.
Dial it in
How many iterations. How many in parallel. At what latency you call it "slow". We pre-fill sensible defaults — you probably don't need to touch them.
Fire it
Tiles land live in a grid as responses come back. At the end, a letter grade, a p95, a slowness ladder — everything you need to say "this is fine" or "this is not fine".
What makes it good.
A short, opinionated list of things that turn a "load test" into a tool people actually open.
Owned hosts only
Pulse refuses to fire at anything outside your environments. No SSRF, no "let me just ping google".
Reuses your bricks
Pick any HTTP brick from your workspace. Pulse copies its URL, method, headers, and auth profile.
Slow is a first-class citizen
Set a threshold in milliseconds. Every slow response shakes, gets a snail, and bumps the slow counter.
Live tiles, not live logs
You don't read lines. You glance at a grid, see the colour distribution, and instantly know.
p50 / p95 / p99
We compute the percentiles every grown-up engineer expects — plus a letter grade for the rest of the team.
Safe by default
Hard caps on iterations & concurrency, per-org in-flight limits, timeout per request, confirm-the-host gate before fire.
Pulse vs. a real load test
Pulse is deliberately tiny. Here's where it fits.
Use Pulse when you want…
- To know if one endpoint feels slow, right now, in 20 seconds.
- A share-able link with a live grid to send to your team.
- A sanity check before a release, without leaving TestingForge.
- A screenshot for the chat that says "look, p95 = 180ms".
- A delight a manual QA can actually run themselves.
Use k6 / Locust / Gatling when you want…
- Tens of thousands of RPS sustained for 30 minutes.
- Complex multi-step scenarios with state and data.
- Long-term regression alerting on latency SLOs.
- A test plan your SRE team lives inside.
Your endpoints deserve
a better smoke test.
Give it 30 seconds. Pick an endpoint, set ×50, and see what falls out. Worst case: you confirm it's fine. Best case: you found a surprise before your users did.
🎁 Launch offer: 30 days free for every TestingForge subscriber.