Getting Started
This page walks you through account setup and running your very first load test.
Step 1 — Create your account
- Go to app.loadcurl.com/auth/register.
- Sign up with your email address.
- Verify your email and log in.
You will land on the Dashboard, which shows your test history, usage quota, and quick-start options.
Step 2 — Paste your endpoint
Click New Test in the top navigation.
In the Endpoint field, enter the full URL of the API you want to test:
https://api.yourapp.com/v2/users
Add authentication headers (optional)
If your endpoint requires authentication, expand the Headers section and add them:
| Header | Example value |
|---|---|
Authorization | Bearer eyJhbGciOiJIUzI1NiIs... |
x-api-key | sk-prod-abc123 |
Content-Type | application/json |
Add a request body (optional)
For POST or PUT endpoints, select the HTTP method from the dropdown and paste your JSON body:
{
"user_id": "demo",
"page": 1,
"limit": 20
}
:::info Import from Postman or OpenAPI
You can skip manual entry entirely. Click Import and upload a Postman collection (.json) or an OpenAPI spec (.yaml / .json) — Load Curl will pre-fill the endpoint, headers, and body for you.
:::
Step 3 — Configure your load profile
The Load Profile panel controls how workers behave during the test.
| Setting | What it does | Starter default |
|---|---|---|
| Concurrent users | Total number of workers running at peak | 50 |
| Duration | How long the test runs at full load | 60 s |
| Ramp-up time | Seconds to gradually reach full concurrency | 10 s |
| Target region | Geographic source of worker traffic | Auto |
Choosing the right concurrency
A common starting point is 10 × your expected peak traffic. For example, if you expect 100 simultaneous users on launch day, start with a 1,000-worker test to build in a safety margin.
Always use a ramp-up period. Sending full concurrency instantly can cause artificial timeouts that hide real bottlenecks.
Step 4 — Run the test
Click Start Test. You will be taken to the Live Dashboard where you can watch metrics stream in real time:
- RPS — current requests per second
- P50 / P95 / P99 latency — response time percentiles updating every second
- Error rate — percentage of non-2xx responses
- Throughput — data transferred per second
- Server CPU — resource usage (when your server agent is installed)
The test runs for the full duration you configured, then finalises the Report Card.
What's next?
- Learn how to read your Report Card
- Build a multi-endpoint scenario to test a full user journey