syntheticcfo

The Python client, command line and MCP server for synthetic cfo: forensic-grade synthetic ERP data for SAP ECC, SAP S/4HANA and Oracle Cloud, generated entirely from accounting rules, with a labelled fraud answer key.

No real company data is ever used. No language model generates any figure: the engine is deterministic Python, and the same seed and configuration regenerate a package byte for byte.

• Product and free sample: https://app.syntheticcfo.com/sample

• The worked proof (real detectors scored on this data, including the one that fails): https://app.syntheticcfo.com/proof

• How to evaluate it in an afternoon: https://app.syntheticcfo.com/evaluate

• Package on the index: https://pypi.org/project/syntheticcfo/

Install

pip install syntheticcfo

Standard library only, no dependencies. Python 3.8 or newer.

Related MCP server: dumate-supply-chain-mock-api

One command

Create an API key in the app (menu, your email, Account, API access), then:

syntheticcfo generate --key scfo_YOURKEY --platform oracle --industry retail --fraud high --rows 600

It builds the world, waits, and saves the package zip where you stand. Add --lab to also save the JSON-lines lab export, --seed 42 to make the world reproducible on purpose, and --evaluation-mode to withhold the ground-truth marker columns from the shipped data. python -m syntheticcfo does the same thing everywhere.

From Python

from syntheticcfo import Client

c = Client(api_key="scfo_...")            # or set SCFO_API_KEY
job = c.generate(
    platform="SAP_ECC",                   # SAP_ECC | SAP_S4HANA | ORACLE
    industry="MANUFACTURING",
    fraud="medium",                       # none | low | medium | high
    rows=1500,
    modules=["P2P", "O2C", "CE", "R2R"],  # subset of P2P O2C CE R2R HR MM FA
    jurisdiction="DE",
    seed=42,
    wait=True,
)
package = job.download("out/")            # the four-layer package (zip)
lab = job.lab_export("out/")              # JSONL tables + answer_key.jsonl + dataset card
print(job.insights())                     # what was planted, by control

What you get

Every dataset is a complete four-layer package: an executive brief with the whitepaper and articulated financial statements, an audit evidence layer with the forensic report and the ground-truth answer key, a SQLite database with full referential integrity, and source workbooks mirroring the real ERP table architecture. Seven cycles (purchase to pay, order to cash, cash and bank, record to report, payroll, inventory, fixed assets), ten jurisdictions, more than 25 fraud schemes each planted beside innocent look-alikes so a detector can be measurably wrong.

MCP server

mcp_server/ lets Claude, Cursor, VS Code or any MCP client generate datasets, download packages and read answer keys in plain language. Three-step setup in mcp_server/README.md.

Plans

The free Developer plan needs no request: all three platforms, the core four cycles, up to 1,000 rows, 5 datasets a month, with the answer key. Paid plans add the remaining modules, volume, multi-year arcs, the blind evaluation track, the intercompany group and the hosted ERP environment. Plan limits apply exactly as in the app; over-plan requests are refused with an honest message.

Licence

The code in this repository (the client, the command line and the MCP server) is MIT licensed; see LICENSE. The platform, the generated data and their terms of use are separate: https://app.syntheticcfo.com/terms