> ## Documentation Index
> Fetch the complete documentation index at: https://docs.sinkove.com/llms.txt
> Use this file to discover all available pages before exploring further.

# SDK Quick Start

> Create your first AI dataset in 5 minutes with the Sinkove Python SDK

## Prerequisites

<CardGroup cols={3}>
  <Card title="Python 3.12+" icon="python">
    Required for SDK
  </Card>

  <Card title="API Key" icon="key">
    From your dashboard
  </Card>

  <Card title="Organization ID" icon="building">
    Your org UUID
  </Card>
</CardGroup>

<Note>
  API keys and Organization IDs are covered in the [Get Started](/essentials/quickstart) section.
</Note>

## 1. Install the SDK

```bash theme={null}
pip install sinkove-sdk
```

## 2. Set Your API Key

```bash theme={null}
export SINKOVE_API_KEY="your-api-key-here"
```

## 3. Create Your First Dataset

Create `first_dataset.py`:

```python theme={null}
import uuid
from sinkove import Client

# Your IDs (replace with actual values)
ORGANIZATION_ID = uuid.UUID("your-organization-id")
MODEL_ID = uuid.UUID("your-model-id")

# Initialize client
client = Client(ORGANIZATION_ID)

# Create dataset
print("Creating dataset...")
dataset = client.datasets.create(
    model_id=MODEL_ID,
    num_samples=10,
    args={"prompt": "chest x-ray showing pneumonia"}
)

print(f"Dataset created! ID: {dataset.id}")

# Wait and download
print("Waiting for completion...")
dataset.wait()

print("Downloading...")
dataset.download("my_first_dataset.zip", strategy="replace")
print("✓ Complete!")
```

## 4. Run the Script

```bash theme={null}
python first_dataset.py
```

Expected output:

```
Creating dataset...
Dataset created! ID: 123e4567-e89b-12d3-a456-426614174000
Waiting for completion...
Downloading...
✓ Complete!
```

## Understanding the Code

* **Client Initialization**: `Client(ORGANIZATION_ID)` connects using your API key
* **Dataset Creation**: Specifies model, sample count, and generation parameters
* **Waiting**: `dataset.wait()` blocks until generation completes
* **Downloading**: Saves the generated dataset locally

## Common Patterns

```python theme={null}
# Check dataset status
print(f"State: {dataset.state}, Ready: {dataset.ready}")

# Handle existing datasets
dataset = client.datasets.get(uuid.UUID("existing-dataset-id"))
if dataset.ready:
    dataset.download("dataset.zip")

# List all datasets
datasets = client.datasets.list()
for ds in datasets:
    print(f"{ds.id}: {ds.state}")
```

## Error Handling

```python theme={null}
try:
    client = Client(ORGANIZATION_ID)
    dataset = client.datasets.create(MODEL_ID, 10, {"prompt": "test"})
    dataset.wait(timeout=300)
    dataset.download("output.zip")
except ValueError as e:
    print(f"Configuration error: {e}")
except TimeoutError:
    print("Dataset generation took too long")
except Exception as e:
    print(f"Unexpected error: {e}")
```

## Next Steps

<CardGroup cols={2}>
  <Card title="Python SDK Guide" icon="python" href="/essentials/python-sdk">
    Complete SDK documentation and advanced features
  </Card>

  <Card title="Code Examples" icon="code" href="/essentials/sdk-examples">
    Advanced patterns and real-world use cases
  </Card>

  <Card title="API Reference" icon="book" href="/essentials/sdk-reference">
    Complete method and class documentation
  </Card>

  <Card title="Model Catalog" icon="cube" href="https://cloud.sinkove.com/models">
    Browse available AI models
  </Card>
</CardGroup>

## Quick Reference

```python theme={null}
# Essential methods
dataset = client.datasets.create(model_id, num_samples, args)
dataset = client.datasets.get(dataset_id)
datasets = client.datasets.list()
dataset.download("output.zip", strategy="replace", wait=True)
dataset.wait(timeout=600)

# Check status
dataset.ready  # bool
dataset.state  # "PENDING", "STARTED", "READY", "FAILED"
```