# Components

Components are the building blocks of a flow. Langflow ships with a wide catalog covering inputs, models, memory, retrieval, tools, and outputs. On DKubeX, additional **DKubeX Providers** components are available that connect directly to the cluster-local SecureLLM service.

## Categories

| Category | What's inside |
|----------|--------------|
| **Input & Output** | Chat input, text input, file loaders, chat output, text output |
| **Data Sources** | Database connectors, API request components, web scrapers |    
| **Models & Agents** | OpenAI, Anthropic, Mistral, Ollama, Hugging Face, and more |
| **LLM Operations** | Prompt templates, chains, and LLM utility components |
| **Files & Knowledge** | File loaders, document parsers, knowledge base connectors |
| **Processing** | Text splitters, parsers, transformers |
| **Flow Control** | Conditional routers, loops, and parallel execution |
| **Utilities** | Python REPL, custom tools, web search |
| **Prototypes** | Experimental components |
| **Tools** | Agents with tool-calling capabilities |
| **DKubeX Providers** | DKubeX-native LLM and embedding components backed by SecureLLM |


![Component categories](./media/Langflow-component-categories.png)

## DKubeX Providers

These components are built into every DKubeX Langflow instance and connect to the **SecureLLM** service running in the same cluster. They appear in the sidebar under **DKubeX Providers**.

![DKubeX components](./media/Langflow-dkubex-providers.png)

### DKubeX LLM

Generates text using language models served by SecureLLM.

![DKubeX LLM component](./media/Langflow-dkubex-llm.png)

**Configuration:**

| Field | Description |
|-------|-------------|
| SecureLLM API Key | Your API key for the SecureLLM service (masked; stored once) |
| Model Name | Dropdown of available `TextGeneration` models — click the refresh button to populate |
| Max Tokens | Maximum tokens to generate (default 256) |
| Temperature | Sampling temperature 0–1 (default 0.1) |
| JSON Mode | Force JSON output format |
| Timeout | Request timeout in seconds (default 700) |
| Max Retries | Retries on transient failure (default 5) |

**How to use:**

1. Drop **DKubeX LLM** component onto the canvas.
2. Enter your **SecureLLM API Key**.
3. Click on **Model Name** and then click **Refresh list**  to load the available models.
4. Select a model and connect the component to your flow.
5. Click the component to open its side panel and configure the advanced options.

### DKubeX Embeddings

Generates vector embeddings using embedding models served by SecureLLM. Use this component with a Vector Store component to index and retrieve documents.


![DKubeX Embeddings component](./media/Langflow-dkubex-embeddings.png)

**Configuration:**

| Field | Description |
|-------|-------------|
| SecureLLM API Key | Your API key for the SecureLLM service (masked; stored once) |
| Model Name | Dropdown of available `TextEmbedding` models — click refresh to populate |
| Chunk Size | Number of texts per embedding batch (default 256) |
| Request Timeout | Timeout in seconds |
| Max Retries | Retries on failure (default 3) |

**How to use:**

1. Drop **DKubeX Embeddings** onto the canvas.
2. Enter your **SecureLLM API Key**.
3. Click on **Model Name** and then click **Refresh list**  to load the available models.
4. Select a model and connect the component to a Vector Store component to index documents.
5. Click the component to open its side panel and configure the advanced options.

> **Note:** The model dropdowns in DKubeX components show only models whose type matches the component — text generation models for **DKubeX LLM** and embedding models for **DKubeX Embeddings**. If the dropdown is empty after refreshing, verify your API key or contact your cluster administrator.

## Configuring Any Component

1. Fill in required fields (marked with `*`).
2. Use the **Code** tab to view or customize the underlying Python code.
3. Click a component on the canvas to open its side panel, containing the advanced options.

## Custom Components

You can create your own component by either modifying the code of an existing component, or by clicking on **New Custom Component** to build a custom component from scratch.