A profile from a recording
The flow you recorded for regression becomes the load profile: same steps, same auth, scaled to virtual users.
Read morePerformance and load · updated 2026-10-05
Performance and load testing
Load profiles from a recording or plain words, run within an enforced virtual-user cap, P50 to P99.9 and Apdex live, export to k6, JMeter or Locust.
For: Performance engineers and QA teams who need load evidence before a release, and teams replacing JMeter scripts nobody wants to maintain.
Capabilities
Intent-based load
Describe the traffic in plain words, such as a launch-day peak, and the platform builds the profile.
Read moreJMeter-equivalent engine
Thread groups, timers and assertions, with the semantics teams already know and without hand-edited XML.
Read morePercentiles and Apdex, live
P50, P95, P99, P99.9, throughput, error rate and Apdex while the run is in progress.
Export to open tools
Every profile exports to k6, JMeter or Locust, so the load test is yours to keep.
Caps that are enforced, not printed
Virtual-user caps per plan are enforced by the platform on every run start.
How it works
- Choose the sourceA recorded flow, an API collection, or a plain-language description of the traffic.
- Shape the loadVirtual users, ramp, duration, think time; the cap for your plan is shown before you start.
- Run and watchPercentiles, throughput and errors update in real time.
- Compare and exportCompare against a previous run, fail a pipeline on a regression, export the profile.
What you receive, and what we do not do
Evidence ledger
- Per run
- P50, P95, P99, P99.9, throughput, error rate and Apdex, plus a per-request table.
- Per profile
- The source flow, the load shape, and the k6, JMeter or Locust export.
- Per plan
- The enforced virtual-user cap: 1,000 on Solo Pro, 10,000 on Team, 30,000 on Enterprise.
Limits we state
- Load testing is not included on the Free plan.
- Load is generated from the platform’s servers in Mumbai; latency includes the distance to your target.
- A guide on performance baselines and SLA validation is planned, not published.
Plans: Solo Pro (1,000 virtual users per run), Team (10,000) and Enterprise (30,000). Not included on Free. See the metering table.
Questions buyers ask
How many virtual users can I run?
The cap is enforced per plan on every run start: 1,000 on Solo Pro, 10,000 on Team, 30,000 on Enterprise, which is also the engine’s own hard bound. The Free plan does not include load testing.
Can I keep using my JMeter scripts?
Yes, in both directions: the migration importer brings JMeter assets in, and every profile built here exports to JMeter, k6 or Locust.
Why percentiles and not an average?
An average hides the slow tail that users notice. P95 and P99 show the time 95 and 99 percent of requests beat; Apdex turns the same data into one score against a target you set.
Try it on your own application
The Free plan has no time limit and asks for no card. Or bring one flow from your product and we record it live.