> For clean Markdown of any page, append .md to the page URL.
> For a complete documentation index, see https://docs.nvidia.com/nemo-helix/v0.7.1/documentation/design-synthetic-data/llms.txt.
> For AI client integration (Claude Code, Cursor, etc.), connect to the MCP server at https://docs.nvidia.com/nemo-helix/_mcp/server.

# Data Designer

Data Designer on NeMo Helix enables high-quality synthetic data generation through the NeMo Data Designer plugin. You can submit workloads to a running NeMo Services cluster using the CLI or SDK.

## Overview

Data Designer is a framework for orchestrating complex synthetic data generation workflows. It coordinates LLM calls, manages dependencies between data fields, handles batching and parallelization, and validates generated data against specifications.

The plugin is built on the open-source [NVIDIA NeMo Data Designer library](https://docs.nvidia.com/nemo/datadesigner/v0.9.3/getting-started/welcome) ([GitHub](https://github.com/NVIDIA-NeMo/DataDesigner)). The library provides the configuration and generation engine; the plugin provides CLI, SDK, Data Designer API, Jobs, Files API, Secrets API, and Inference Gateway API integration.

## How It Works

Data Designer separates **configuration** from **execution**.

### 1. Build Configurations

Use `data_designer.config` to define the dataset you want to generate:

```python
import data_designer.config as dd

# Define models
model_configs = [
    dd.ModelConfig(
        provider="default/nvidia-build",
        model="nvidia/nemotron-3.5-lightning-30b-a3b",
        alias="text",
    )
]

# Build configuration
config_builder = dd.DataDesignerConfigBuilder(model_configs)
config_builder.add_column(
    dd.SamplerColumnConfig(
        name="sentiment",
        sampler_type=dd.SamplerType.CATEGORY,
        params=dd.CategorySamplerParams(values=["positive", "negative"]),
    )
)
config_builder.add_column(
    dd.LLMTextColumnConfig(
        name="review",
        prompt="Write one short " + "{0}{0} sentiment {1}{1}".format("{", "}") + " product review.",
        model_alias="text",
    )
)
```

Configuration code describes the dataset schema, columns, dependencies, constraints, seed data, processors, profilers, and inference settings.

**Learn more**: See the [library documentation](https://docs.nvidia.com/nemo/datadesigner/v0.9.3/getting-started/welcome) for comprehensive guides on column types, samplers, constraints, and advanced features.

### 2. Execute on the Platform

The dataset defined by your configuration is executed on NeMo Helix, integrating with these NeMo Services APIs:

| Integration               | What it provides                                                    |
| ------------------------- | ------------------------------------------------------------------- |
| **Inference Gateway API** | Centralized model providers and OpenAI-compatible inference routes. |
| **Files API**             | Filesets for seed data and persona datasets.                        |
| **Secrets API**           | API keys and tokens referenced from Data Designer configurations.   |
| **Jobs API**              | Service-managed create workloads, logs, status, and artifacts.      |

## Next Steps

#### [CLI](/documentation/design-synthetic-data/cli)

Run previews and create datasets with `nemo data-designer`.

#### [Retrieval SDG](/documentation/design-synthetic-data/retrieval-sdg)

Replicate Nemotron embed/rerank Stage 0 and Stage 1 on the platform.

#### [Tutorials](/documentation/design-synthetic-data/tutorials)

Learn through examples: basics, seeding, and more.

#### [Migration Guide](/documentation/design-synthetic-data/migrating-from-standalone-library)

Learn about differences between the standalone library and the NeMo Helix plugin.

#### [Library Documentation](https://docs.nvidia.com/nemo/datadesigner/v0.9.3/getting-started/welcome)

Comprehensive guides on column types, constraints, and advanced features.

## Docs

- [Data Designer CLI](https://docs.nvidia.com/nemo-helix/v0.7.1/documentation/design-synthetic-data/cli.md): 
- [Retrieval SDG](https://docs.nvidia.com/nemo-helix/v0.7.1/documentation/design-synthetic-data/retrieval-sdg.md): Generate and prepare retrieval training data with the Nemotron Stage 0 and Stage 1 recipes.
- [Tutorials](https://docs.nvidia.com/nemo-helix/v0.7.1/documentation/design-synthetic-data/tutorials.md): 
- [The Basics](https://docs.nvidia.com/nemo-helix/v0.7.1/documentation/design-synthetic-data/tutorials/the-basics.md): 
- [Seeding with External Datasets](https://docs.nvidia.com/nemo-helix/v0.7.1/documentation/design-synthetic-data/tutorials/seeding-with-external-datasets.md): 
- [Generate Retrieval Training Data](https://docs.nvidia.com/nemo-helix/v0.7.1/documentation/design-synthetic-data/tutorials/retrieval-generate.md): 
- [Data Designer SDK Resources](https://docs.nvidia.com/nemo-helix/v0.7.1/documentation/design-synthetic-data/sdk-resources.md): 
- [Migrating from Standalone Library](https://docs.nvidia.com/nemo-helix/v0.7.1/documentation/design-synthetic-data/migrating-from-standalone-library.md):