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

# useCreateSamplingMessage

> Request LLM completions from the host.

<Badge color="yellow">sunpeak API</Badge>

## Overview

Returns a function to request LLM completions from the host via MCP `sampling/createMessage`. The host decides which model to use and may modify or reject the request.

Check `getHostCapabilities()?.sampling` before calling. For tool-augmented completions, check `?.sampling?.tools`.

## Import

```tsx theme={null}
import { useCreateSamplingMessage } from 'sunpeak';
```

## Signature

```tsx theme={null}
function useCreateSamplingMessage(): (
  params: CreateSamplingMessageParams
) => Promise<CreateMessageResult | CreateMessageResultWithTools | undefined>
```

### CreateSamplingMessageParams

Standard MCP `CreateMessageRequest` params:

<ResponseField name="messages" type="SamplingMessage[]" required>
  Conversation messages to send to the model.
</ResponseField>

<ResponseField name="maxTokens" type="number" required>
  Maximum tokens in the response.
</ResponseField>

<ResponseField name="systemPrompt" type="string">
  System prompt for the model.
</ResponseField>

<ResponseField name="temperature" type="number">
  Sampling temperature.
</ResponseField>

<ResponseField name="tools" type="Tool[]">
  Tools the model can call. When provided, the result may include `tool_use` blocks.
</ResponseField>

## Returns

<ResponseField name="createSamplingMessage" type="(params) => Promise<CreateMessageResult | CreateMessageResultWithTools | undefined>">
  Function to request an LLM completion.
</ResponseField>

### CreateMessageResult

<ResponseField name="model" type="string">
  Model used by the host.
</ResponseField>

<ResponseField name="role" type="string">
  Response role (typically `"assistant"`).
</ResponseField>

<ResponseField name="content" type="Content">
  Response content block.
</ResponseField>

<ResponseField name="stopReason" type="string">
  Why the model stopped (`"endTurn"`, `"toolUse"`, etc.).
</ResponseField>

## Usage

```tsx theme={null}
import { useCreateSamplingMessage, useHostInfo } from 'sunpeak';

function SummaryButton({ text }: { text: string }) {
  const createSamplingMessage = useCreateSamplingMessage();
  const { hostCapabilities } = useHostInfo();

  if (!hostCapabilities?.sampling) return null;

  const handleSummarize = async () => {
    const result = await createSamplingMessage({
      messages: [
        { role: "user", content: { type: "text", text: `Summarize: ${text}` } },
      ],
      maxTokens: 200,
    });
    if (result) {
      console.log("Summary:", result.content);
    }
  };

  return <button onClick={handleSummarize}>Summarize</button>;
}
```