Python SDK

The official Python SDK is automators-com/datamaker-py. Its import name is datamaker.

Install

pip install "git+https://github.com/automators-com/datamaker-py.git"

For a reproducible environment, pin the Git URL to a reviewed commit. Do not install the unrelated bare datamaker package from PyPI.

Authenticate

DataMaker() reads DATAMAKER_API_KEY and DATAMAKER_API_URL. The default API URL is https://api.datamaker.automators.com. You can pass api_key and base_url explicitly. Use a project-scoped key, or pass scope headers when needed:

import os
from datamaker import DataMaker

dm = DataMaker(default_headers={
    "X-Project-Id": os.environ["DATAMAKER_PROJECT_ID"],
})

Keep API keys in your environment or secret manager. Inside a scenario, the runner injects the API credentials and run context; DataMaker() uses those settings.

Define a template and generate

from datamaker import DataMaker, Template

dm = DataMaker()
template = Template(name="Customer", quantity=3, fields=[
    {"name": "first_name", "type": "First Name"},
    {"name": "last_name", "type": "Last Name"},
    {"name": "email", "type": "Derived",
     "options": {"value": "{{first_name}}.{{last_name}}@example.com"}},
])
result = dm.generate(template)

For a saved template:

result = dm.generate_from_template_id("<template-id>", quantity=100)

Both methods return the parsed API object with live_data and dependencies. They do not return a list of rows.

Convert columns to rows

def to_rows(result):
    columns = result["live_data"]
    if not columns:
        return []
    names = list(columns)
    lengths = {len(columns[name]) for name in names}
    if len(lengths) != 1:
        raise ValueError("Generation returned columns of different lengths")
    return [dict(zip(names, values)) for values in zip(*(columns[n] for n in names))]

rows = to_rows(result)
print(f"Generated {len(rows)} rows")

Inspect the rows for generation errors before exporting. Successful HTTP status alone does not establish that every field contains usable data.

Work with resources

The Python client exposes flat methods such as get_templates(), get_connections(), get_endpoints() and get_projects(). Typed dictionaries describe responses; they remain ordinary Python dictionaries at runtime.

Consult the SDK source for exact method arguments, especially write payloads. Use the REST API for operations not wrapped by your installed SDK version.

Errors

from datamaker.error import DataMakerError

try:
    result = dm.generate_from_template_id("<template-id>", quantity=10)
except DataMakerError as error:
    raise RuntimeError("Data generation failed; inspect the API response") from error

Avoid printing exceptions containing sensitive upstream payloads in shared logs. For file outputs, use the scenario workspace directories.