synthetic-data-generator

Why we included this project

Teams that need realistic tabular data without exposing real records will find a strong fit here. SDG bundles single-table and multi-table synthesis algorithms alongside LLM-driven generation, so you can reach for a classical model when you want speed or lean on a language model when your schemas and relationships get intricate. It also handles the unglamorous parts, like billion-row-scale workloads and privacy-oriented generation, which matters when the output will be shared across teams or used to drive downstream testing and analytics. The project ships as a Python package with Colab examples and API documentation, so it works for a quick experiment and for more deliberate pipelines alike. If you have been hand-rolling fake data or juggling several disjoint generators, this is a more coherent alternative worth evaluating for your own data needs.