Bring a test environment to life.
Seed related tables, exercise edge cases and reproduce bugs with a fixed setup. Export rows in a format your existing tools can use.
Explore schema dataThe ScenarioLoom field guideData with context.
Data with context.
Scenarios with answers.
From a developer’s first seed file to a financial crime testing pack, turn connected synthetic records into a repeatable test you can understand.
No account. No generation API. Your browser does the work.
Input to your screening system
- payment_id
- PAY-042
- creditor_name
- Orchard Trading TEST
- currency
- HKD
↳ Separate answer key
- field
- creditor_name
- expected_hit
- true
- case
- Planted name match
Illustrative fixture · inputs and labels stay separate
Five ways to buildSchema data + four financial crime modes
Repeatable by designSeed, settings, date and engine versions
On your deviceBrowser processing + an offline CLI
Built for the work you doA useful starting point
A useful starting point
for your next test.
Use synthetic records to exercise a workflow, compare detection logic or reproduce a result without bringing real customer data into the exercise.
Know what your rules should find.
Mix planted patterns with legitimate controls. Use the separate labels to understand missed cases and false positives under the supplied fixture assumptions.
See the testing workflowKeep evaluation repeatable.
Save a versioned setup, run the same pack through a new integration, and compare results. Use the CLI to generate fixtures in a local or CI workflow.
Explore local & CLI useOne workspace, five modesChoose the question
Choose the question
you want to test.
Start with a schema, or choose a curated scenario pack. The generator gives you the source records and, for scenario modes, the expected results.
Rows that fit together.
Start with SQL, TypeScript or a JSON sample. Generate related records with consistent identities and relationships, ready for development and QA.
- Keep foreign keys connected across tables.
- Adjust row counts, countries and column generators.
- Exercise nulls, boundaries and unusual text with edge-case mode.
Give your monitoring rules a controlled challenge.
Build a synthetic bank with everyday activity, planted laundering patterns and legitimate look-alikes. Compare your own alerts with a separate answer key.
- Test structuring, layering, mule networks and other patterns at different difficulties.
- Keep ordinary customers and planted scenarios in the same population.
- Import JSON/CSV alerts locally and measure detection and false positives.
Plant the match. Keep the control.
Create fictional watchlists and payments, then test how your screening system handles exact identifiers, aliases, spelling changes and near misses.
- Place matches in party names, bank BICs, remittance text, cities and vessel identifiers.
- Exercise accents, word order, spacing and paired Chinese name variants.
- Export OFAC-style list files and base pacs.008 payment XML.
Follow the relationship to the expected outcome.
Explore curated company networks with ownership stakes, control rights, dated designations and payment records. The answer key distinguishes named US, EU and UK test profiles.
- Test indirect chains, independent owners, exact thresholds and explicit joint arrangements.
- Inspect the graph and the ownership/control evidence behind it.
- Keep incomplete post-designation transfer facts in a review category.
Put the documents in context.
Link invoices, shipments, structured credit records and financing draws. Test pricing, business scope, duplicate financing, routing and fictional screening restrictions.
- Compare invoice prices with the supplied illustrative references.
- Distinguish duplicate full-value draws from documented installments.
- Use premium-grade certificates and route explanations as legitimate controls.
Your first ten minutes01 02 03 04
Build. Inspect. Export. Compare.
Choose your inputs
Open a mode. Paste a supported schema or choose scenario types, volume and countries.
Fix the setup
Set a seed and anchor date. Small requests update as you edit; use Generate for larger packs. Cancel whenever you need.
Review both sides
Inspect source records and expected outcomes. Keep answer-key tables away from the system under test.
Run your comparison
Export the inputs. Save the setup and labels. In AML mode, import your system’s alerts for a local score report.
Need to repeat the run? Use Download setup to save a versioned recipe. A share link also carries the active setup; read what it contains before sharing it.
Use-case guidesStart from a
Start from a
ready-made setup.
Each guide explains one kind of test data, shows real generated output and opens the matching setup in the workspace.
AML TRANSACTION MONITORING
Synthetic AML transaction monitoring test data with answer keys
Generate a fictional bank with everyday activity, planted money-laundering typologies, legitimate look-alikes and a separate answer key.
SANCTIONS SCREENINGSanctions screening test data
Generate a fictional sanctions watchlist and payments with planted name matches, typos, aliases, transliterations and near-miss controls.
ISO 20022 PAYMENTSISO 20022 pacs.008 sample XML generator for testing
Generate fictional ISO 20022 pacs.008.001.08 customer credit transfer messages with realistic parties, BICs and remittance, plus planted sanctions-screening hits and a separate answer key..
OWNERSHIP AND CONTROLBeneficial ownership (UBO) and control test data
Generate fictional ownership networks with direct, aggregated, indirect and board control cases, dated designations and payments, plus an answer key under named US, EU and UK test profiles..
TRADE-BASED MONEY LAUNDERINGTrade-based money laundering (TBML) test scenarios
Generate fictional trade-finance cases: over- and under-invoicing, duplicate financing, unexplained routing, listed vessels and ports, with legitimate controls and an answer key.
SCHEMA TO SEED DATAGenerate realistic seed data from a SQL or TypeScript schema
Paste a PostgreSQL, MySQL, SQLite or SQL Server schema, TypeScript interfaces or a JSON sample and get related, reproducible seed data: valid foreign keys, consistent order totals, JSON, CSV or SQL INSERTs..
The answer key mattersGive every planted case
Give every planted case
a clear expectation.
A useful benchmark includes the pattern, the legitimate look-alike and the reason each has its expected outcome.
ScenarioLoom keeps those labels separate from source records. Ownership and trade fixtures can also request review when the supplied facts do not support a definite result.
Test fixtures, with a defined scope. These are fictional scenarios and illustrative parameters. They do not assess real companies or authorize payments. Ownership guidance profiles are dated; XML targets the documented base schemas; trade credit records are structured fixtures, not SWIFT MT700 instruments. Use an isolated test environment and get practitioner review before making a compliance claim.
Make the next test repeatable.
Start in the browser, or take a copy into your own environment.