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

# Linkup

> Linkup is an AI dev tool that helps you connect your code with your LLM. It’s a platform that allows you to connect your code with your LLM and get observability.

Check out the [LinkUp website](https://www.linkup.so/) for more information.

## 1. Add Linkup Parameters to Your Request

Include `linkup_params` in your request's `keywordsai_params`:

```json {6-12} theme={"system"}
{
  "messages": [
    {"role": "system", "content": "You are a helpful assistant."},
    {"role": "user", "content": "Please summarize the information about Microsoft's 2024 revenue: {% if linkup_search_response %}{{ linkup_search_response.answer }}{% endif %}"}
  ],
  "linkup_params": {
      "api_key": {{"YOUR_LINKUP_API_KEY"}},
      "q": "What is Microsoft's 2024 revenue?",
      "depth": "deep",
      "outputType": "sourcedAnswer",
      "includeImages": false
    }
  }
```

## 2. Use Template Variables in Your Messages

The Linkup response will be available as a variable named `linkup_search_response` in your templates. You can access it using Jinja2 syntax:

<Tip>
  You can learn how to use Jinja2 syntax in the [Jinja2 documentation](/documentation/products/prompt_management/creating_prompts/jinja_support).
</Tip>

```
{% if linkup_search_response %}
  {{ linkup_search_response.answer }}
  
  Sources:
  {% for source in linkup_search_response.sources %}
    - {{ source.name }}: {{ source.url }}
      {{ source.snippet }}
  {% endfor %}
{% endif %}
```

## 3. Using with Prompts

If you're using [Keywords AI Prompts](/documentation/products/prompt_management/quickstart), you can include the Linkup template variables in your prompt templates. When you make a request with `linkup_params`, the variables will be populated and rendered in your prompt.

## How It Works

1. When a request with `linkup_params` is received, the system:
   * Extracts the Linkup parameters from the request
   * Calls the Linkup API to retrieve search results
   * Adds the results to the template variables as `linkup_search_response`
   * Renders the prompt templates with these variables

2. If you're using stored prompts, the system:
   * Loads the prompt without rendering variables
   * Merges variables from the prompt and the request
   * Processes Linkup parameters and adds results to variables
   * Renders the prompt with all variables, including Linkup results

## Parameters

The `linkup_params` object supports the following parameters:

| Parameter                | Type    | Description                                                                 | Default           |
| ------------------------ | ------- | --------------------------------------------------------------------------- | ----------------- |
| `q`                      | string  | The search query                                                            | Required          |
| `depth`                  | string  | The depth of the search (`"standard"` or `"deep"`)                          | `"deep"`          |
| `outputType`             | string  | The type of output (`"searchResults"`, `"sourcedAnswer"` or `"structured"`) | `"sourcedAnswer"` |
| `structuredOutputSchema` | string  | Schema for structured output                                                | `null`            |
| `includeImages`          | boolean | Whether to include images in the response                                   | `false`           |

## Response Structure

The `linkup_search_response` variable will contain:

```json theme={"system"}
{
  "answer": "The answer text",
  "sources": [
    {
      "name": "Source name",
      "url": "Source URL",
      "snippet": "Text snippet from the source"
    },
   // ...
  ]
}
```

## Example

Here's a complete example of using the Linkup integration:

```python Python theme={"system"}
import requests

def demo_call(input_text, 
              model="gpt-4o-mini",
              token="YOUR_KEYWORDS_AI_API_KEY"):
    headers = {
        'Content-Type': 'application/json',
        'Authorization': f'Bearer {token}',
    }

    data = {
        "messages": [
            {"role": "system", "content": "You are a helpful assistant."},
            {"role": "user", "content": input_text}
        ],
        "linkup_params": {
            "q": "What is Microsoft's 2024 revenue?",
            "depth": "deep",
            "outputType": "sourcedAnswer",
            "includeImages": False
        },
        "variables": {
            "company_name": "Microsoft"
        },
        "model": model
    }

    response = requests.post('https://api.keywordsai.co/api/chat/completions', headers=headers, json=data)
    return response

input_text = "Please provide information about {{ company_name }}'s 2024 revenue and cite your sources."
response = demo_call(input_text)
print(response.json())
```

With a prompt template like:

```
Please provide information about {{ company_name }}'s 2024 revenue and cite your sources.

{% if linkup_search_response %}
Here's what I found:
{{ linkup_search_response.answer }}

Sources:
{% for source in linkup_search_response.sources %}
- {{ source.name }}: {{ source.url }}
{% endfor %}
{% endif %}
```

## Error Handling

If there's an error with the Linkup API call, it will be added to the `warnings_dict` in the response with the key `LINKUP_SEARCH_ERROR`.