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

# Client

`Client` is the main class exposed by CyborgDB. It exposes the functionality necessary to create, load, list and delete indexes. Operations within encrypted indexes (such as `upsert` and `query`) are contained within the `EncryptedIndex` class returned by `create_index` and `load_index`.

## Constructor

```python theme={null}
Client(api_key: str = "",
       storage_config: StorageConfig,
       cpu_threads: int = 0,
       gpu_config: GPUConfig | None = None)
```

Initializes a new CyborgDB `Client` instance.

### Parameters

| Parameter        | Type                                        | Default | Description                                                                                                                                                                                                                                                                                    |
| ---------------- | ------------------------------------------- | ------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| `api_key`        | `str`                                       | `""`    | *(Optional)* API key for your CyborgDB account. If omitted or empty, the client runs in free-tier mode, capped at 1,000,000 items per index. Pass a key from the [Admin Dashboard](../../../intro/get-api-key) for unlimited usage. When omitted, pass `storage_config` as a keyword argument. |
| `storage_config` | [`StorageConfig`](../types#storageconfig)   | -       | Backing store for all index keystores. Built via the static factories `StorageConfig.memory()`, `StorageConfig.disk(path)`, or `StorageConfig.s3(bucket)`.                                                                                                                                     |
| `cpu_threads`    | `int`                                       | `0`     | *(Optional)* Number of CPU threads to use for computations (defaults to `0` = all cores).                                                                                                                                                                                                      |
| `gpu_config`     | [`GPUConfig`](../types#gpuconfig) \| `None` | `None`  | *(Optional)* GPU operations configuration. Specify which operations use GPU acceleration (defaults to `None`, no GPU).                                                                                                                                                                         |

<Note>A single `storage_config` replaces the former `index_location` / `config_location` / `items_location` trio. The store is shared across all per-index keystores.</Note>

### Exceptions

<AccordionGroup>
  <Accordion title="ValueError">
    * Throws if the `cpu_threads` parameter is less than `0`.
    * Throws if the [`StorageConfig`](../types#storageconfig) is invalid.
    * Throws if the GPU is not available when `gpu_config` is set.
  </Accordion>

  <Accordion title="RuntimeError">
    * Throws if the backing store is not available.
    * Throws if the Client could not be initialized.
  </Accordion>
</AccordionGroup>

### Example Usage

```python theme={null}
import cyborgdb_core as cyborgdb

# Example 1: In-memory backing store, free-tier (no API key)
client1 = cyborgdb.Client(
    storage_config=cyborgdb.StorageConfig.memory(),
    cpu_threads=4
)

# Example 2: With an API key for unlimited usage, GPU on upsert and train
api_key = "your_api_key_here"  # Replace with your CyborgDB API key
gpu_config = cyborgdb.GPUConfig(upsert=True, train=True)
client2 = cyborgdb.Client(
    api_key=api_key,
    storage_config=cyborgdb.StorageConfig.disk("/tmp/cyborgdb"),
    cpu_threads=4,
    gpu_config=gpu_config
)

# Proceed with further operations
```

<Tip>Use `StorageConfig.memory()` for ephemeral testing, `StorageConfig.disk("/path")` for local persistence, or `StorageConfig.s3("bucket")` for S3-backed persistence. See [`StorageConfig`](../types#storageconfig).</Tip>

***

## Methods

### get\_cpu\_threads()

Returns the number of CPU threads configured for this client.

```python theme={null}
def get_cpu_threads(self) -> int
```

#### Returns

`int`: The number of CPU threads.

#### Example Usage

```python theme={null}
threads = client.get_cpu_threads()
print(f"Using {threads} CPU threads")
```

***

### is\_gpu\_enabled()

Checks if GPU acceleration is enabled for this client.

```python theme={null}
def is_gpu_enabled(self) -> bool
```

#### Returns

`bool`: True if GPU acceleration is enabled, False otherwise.

#### Example Usage

```python theme={null}
gpu_enabled = client.is_gpu_enabled()
print(f"GPU acceleration is {'enabled' if gpu_enabled else 'disabled'}")
```

***

### get\_gpu\_config()

Returns the GPU operations configuration for this client.

```python theme={null}
def get_gpu_config(self) -> GPUConfig
```

#### Returns

[`GPUConfig`](../types#gpuconfig): The GPU operations configuration.

#### Example Usage

```python theme={null}
gpu_config = client.get_gpu_config()
print(f"GPU operations: upsert={gpu_config.upsert}, train={gpu_config.train}, query={gpu_config.query}")
```