OpenViSTAI · Agents · Blockchain · Crypto
Submit a project
Menu
AI Data & Analytics

Datadesigner

Datadesigner is an open-source Python library for generating synthetic datasets from scratch or seed data, including text, structured data and image outputs, with workflows that can use image, audio and video context.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Datadesigner — OpenViST project profile in AI Data & Analytics.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Use it when a Python-based workflow needs synthetic data generation that can combine seed data, structured sampling, model-generated content or validation. The repository's quick start demonstrates configuring fields and previewing a sample record.

Who it is for

  • AI developers creating synthetic datasets for model development or data augmentation.
  • Python developers who need configurable generation, validation or multimodal data workflows.

Documented uses

  • Create synthetic datasets from scratch or seed data.
  • Augment existing datasets with generated examples.
  • Combine sampled structured fields and LLM-generated text, including fields that depend on other fields.
  • Validate and score generated outputs with Python, SQL, custom validators or LLM judges.
  • Connect generation to local or remote MCP servers and capture interaction traces.
Before you choose

Requirements & limits

The project is public and open source under the Apache-2.0 license. The README lists Python 3.10–3.14 and documents pip or source installation. Its quick start uses an API key from NVIDIA Build, OpenAI or OpenRouter.

  • The documented entry point is a Python configuration builder and library interface, rather than a hosted end-user application.
  • The quick-start workflow requires an API key for a model provider.
  • The project documents text, structured and image generation, plus image, audio and video context in multimodal workflows.
  • Installation downloads third-party open-source projects; the README advises reviewing their licenses.

Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.

Source-based practical guide

Install Data Designer and preview a synthetic review record

Install the Python package, configure a model-provider key, and preview a sample record with a generated product review.

Before you start

  • Python 3.10–3.14, as listed in the repository README.
  • An API key from NVIDIA Build, OpenAI, or OpenRouter for the selected provider.

Steps

  1. Install the package with the command below.

    Documented command
    pip install data-designer

  2. Set the selected provider’s environment variable, such as NVIDIA_API_KEY, OPENAI_API_KEY, or OPENROUTER_API_KEY.

  3. In Python, initialize DataDesigner and DataDesignerConfigBuilder; add a categorical product_category sampler and an LLMTextColumnConfig review column using model_alias="nvidia-text" and a prompt that references {{ product_category }}.

  4. Call data_designer.preview(config_builder=config_builder) and then preview.display_sample_record().

Expected outcome

A displayed sample record preview containing a sampled product category and a generated review, if the configured model provider is available and configured.

Guide limitations

  • Installation also downloads third-party open-source software; review those projects’ license terms before use.
  • The README documents previewing a sample record, not generating or exporting a full dataset.
  • The example uses the nvidia-text model alias; the README does not explain how to configure that alias.
Documentation 1 ↗Documentation 2 ↗
Evidence and transparency

Sources & review

Official repository ↗Guide documentation ↗
Source record details
Source availability
Active
Sources checked
9 October 2026
Original website title
NVIDIA-NeMo/DataDesigner · Official homepage title · retained in its source language
How we review sourcesSuggest a correction