Frequently asked questions

Recipes define generated records: their fields, types, language, quantity, and format. Catalogs provide curated reference lists, such as countries, languages, currencies, names, and cities. Catalog data can be exported directly or used by fields in a recipe.

Create a free account, build and save a recipe or choose a catalog export, review the credit estimate, and start the job. When generation finishes, download the result. The TL;DR gives the shortest walkthrough, and the documentation covers every option.

Recipes can produce CSV, JSON, TSV, XLSX, PDF, XML, YAML, and SQL. Some specialized testing packages and catalog export formats are also available; see the current format list.

The early beta uses conservative per-job limits for records, fields, request size, processing time, and memory. The service validates each request before charging it and explains when a limit is exceeded. Large-volume generation and delivery are planned for a later release; for now, Ghost Bakery is best suited to small and medium test fixtures.

Recipe generation costs one credit per generated field value, plus a 10% processing overhead rounded up. Every field type costs the same. Catalog exports cost one credit per returned row. You see the exact estimate before spending credits. During the start of the beta, free credits are available but purchasing additional credits is not yet available.

Eligible free accounts are topped back up to 500 free credits every 30 days. Autorefill is not cumulative: it restores the balance up to the free allowance instead of adding 500 credits on top of unused credits each period.

Generated records are synthetic and intended for lawful testing. They may resemble real values by chance, but they are not a source of real personal records. Catalogs contain public reference facts and vocabulary from attributed sources. Do not put secrets or real personal data into recipes.

Synthetic output may be used in your lawful software testing under the Ghost Bakery Terms of Use. Catalogs combine public reference sources with their own licenses and attribution requirements; the source and license for each dataset are listed in the data sources section.

Common uses include unit-test fixtures, integration and contract tests, QA and staging data, demos, API payloads, database seeds, Unicode testing, and load-test inputs. Hosted calls are usually best kept in test setup or integration suites; isolated unit tests can consume previously generated fixtures.

No. A recipe preserves the data shape and rules, but generated values are random and can change on each run. Tests should normally assert schemas and invariants rather than exact generated values.

Yes. The HTTP API supports scoped API keys and exposes an interactive API reference. Use the minimum scopes required by your automation and keep keys out of source control and logs.

We are working on public SDKs for Python, JavaScript/TypeScript, and Go. Until they are released, use the documented HTTP API directly.

Email support@ghostbakery.com. Questions, bug reports, generator ideas, language feedback, and suggestions are welcome.