Upload
CSV datasets
Upload any CSV file as a reusable dataset. Choose the delimiter, quoting, encoding, and recycle behavior per column so the runner gets exactly what your script expects.
Test data
Realistic datasets for every load test.
Stop hand-building CSVs. MaxoPerf lets you upload existing datasets or generate synthetic, correlated test data with the full faker catalog — then preview rows live, bind to any test, and choose how data reaches your load generators.
What you get
A single place to manage every dataset your performance tests depend on — from one-off CSVs to rich synthetic catalogs with hundreds of correlated fields.
Capabilities
Everything your tests need, nothing they don't
From simple CSV uploads to multi-field synthetic families, test data management is built into the same workflow as your runs.
Generate
Full synthetic faker catalog
Generate realistic test data from hundreds of built-in generators: names, ages, email addresses, phone numbers, credit-card numbers, airline seat assignments, currency codes, postal addresses, company names, and more — all in one catalog.
Correlate
Correlated, realistic data families
Group fields into families so every row is coherent. A person family keeps first name, last name, email, and address consistent row-to-row. An order family ties product, quantity, and price together. No more mismatched columns.
Preview
Live preview while authoring
See sample rows generated from your exact schema before you save. Catch mismatched types, formatting errors, and family inconsistencies immediately — without running a test.
Distribute
Split across load generators or replicate to every runner
Choose how data reaches your fleet. Split divides the dataset into disjoint slices so each load generator works on unique rows — ideal for user journeys that must not collide. Replicate sends the full dataset to every runner for scenarios that need each worker to have a complete copy.
Compatible
Works with every test engine
Datasets arrive as a local CSV at a stable path inside every runner, regardless of engine. Taurus, JMeter, k6, Gatling, Locust — and all other supported engines — can read the same file path without any script changes.
How it works
From dataset to running test in five steps
The test data workflow is built into the existing run flow — no separate tool to learn.
- Create a dataset. Open the Test data tab in your workspace, name the dataset, and add parameters — each one maps to a CSV column or a synthetic generator field.
- Generate or upload rows. Pick generators from the faker catalog and configure any arguments (locale, format, length), or upload a CSV file. Mix both in the same dataset.
- Keep it realistic. Link fields into a correlated family so each row is coherent — a person record with a matching email and address, a financial record with a consistent card type and currency.
- Preview before you commit. Use the live preview to inspect sample rows and confirm the data looks right. No run needed to catch problems.
- Bind to a test and choose distribution. Attach the dataset to a test configuration. Choose split for disjoint per-runner slices, or replicate to give every runner a full copy. The platform delivers the data automatically at run time.
Use cases
Built for realistic load scenarios
Test data management closes the gap between synthetic load and realistic user behavior.
E-commerce checkout flows
Generate correlated product, user, and payment records so each simulated shopper has consistent cart contents, delivery addresses, and payment details — without reusing the same test account across thousands of virtual users.
Financial and regulated workloads
Produce realistic credit-card numbers, IBAN strings, and transaction amounts for stress tests on payment services. Masked fields keep sensitive-looking values out of logs and run evidence.
Travel and booking systems
Combine airline seat assignments, passenger names, and booking references into coherent reservation records to exercise booking engines without reusing stale fixtures.
Account and authentication tests
Create unique usernames, email addresses, and passwords for sign-up or login tests. Split the dataset so concurrent runners never compete for the same credentials.