> For the complete documentation index, see [llms.txt](https://docs.agenticflow.ai/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.agenticflow.ai/reference/nodes/openai_search.md).

# OpenAI Web Search

**Action ID:** `openai_search`

## Description

Searches the web using OpenAI's capabilities. This node performs real-time web searches integrated with OpenAI models, allowing you to get current information and search results directly within your workflows. Supports batch searching with multiple queries simultaneously.

## Provider

**OpenAI**

## Connection

| Name              | Description                                  | Required | Category |
| ----------------- | -------------------------------------------- | :------: | -------- |
| OpenAI Connection | The OpenAI connection to use for the search. |     ✓    | openai   |

## Input Parameters

| Name            | Type     | Required | Default                          | Description                                                                                                          |
| --------------- | -------- | :------: | -------------------------------- | -------------------------------------------------------------------------------------------------------------------- |
| search\_queries | array    |     ✓    | -                                | List of search queries to process using OpenAI's web search capability. Accepts 1-10 queries.                        |
| model           | dropdown |     -    | gpt-4o-search-preview-2025-03-11 | The OpenAI model to use for search. Options: gpt-4o-search-preview-2025-03-11, gpt-4o-mini-search-preview-2025-03-11 |

<details>

<summary>View JSON Schema</summary>

```json
{
  "description": "OpenAI Search node input.",
  "properties": {
    "search_queries": {
      "description": "List of search queries to process using OpenAI's web search capability",
      "items": {
        "type": "string"
      },
      "maxLength": 10,
      "minLength": 1,
      "title": "Search Queries",
      "type": "array"
    },
    "model": {
      "default": "gpt-4o-search-preview-2025-03-11",
      "description": "The OpenAI model to use for search",
      "enum": [
        "gpt-4o-search-preview-2025-03-11",
        "gpt-4o-mini-search-preview-2025-03-11"
      ],
      "title": "Model Name",
      "type": "string"
    }
  },
  "required": [
    "search_queries"
  ],
  "title": "OpenAISearchInput",
  "type": "object"
}
```

</details>

## Output Parameters

| Name            | Type | Description                     |
| --------------- | ---- | ------------------------------- |
| search\_results | any  | The search results from OpenAI. |

<details>

<summary>View JSON Schema</summary>

```json
{
  "description": "OpenAI Search node output.",
  "properties": {
    "search_results": {
      "title": "Search Results",
      "description": "The search results from OpenAI"
    }
  },
  "title": "OpenAISearchOutput",
  "type": "object"
}
```

</details>

## How It Works

This node integrates OpenAI's web search capabilities directly into your workflows. You provide one or more search queries, and the node performs real-time web searches and returns comprehensive results. The search models (gpt-4o-search-preview or gpt-4o-mini-search-preview) combine language understanding with web search to deliver relevant, up-to-date information. You can search for multiple queries in a single node execution (up to 10 queries), making it efficient for batch processing or comprehensive research tasks.

## Usage Examples

### Example 1: Single Query Search

**Input:**

```
search_queries: ["latest artificial intelligence breakthroughs 2024"]
model: "gpt-4o-search-preview-2025-03-11"
```

**Output:**

```
search_results: {
  "results": [
    {
      "title": "OpenAI Announces GPT-4.1...",
      "url": "https://openai.com/...",
      "snippet": "OpenAI announced GPT-4.1 with improved reasoning capabilities..."
    },
    {
      "title": "Google Releases Gemini 2.0...",
      "url": "https://google.com/...",
      "snippet": "Google's new Gemini 2.0 model demonstrates advanced multimodal capabilities..."
    }
  ]
}
```

### Example 2: Multiple Query Search

**Input:**

```
search_queries: [
  "Python web scraping libraries",
  "FastAPI best practices",
  "Redis caching strategies"
]
model: "gpt-4o-search-preview-2025-03-11"
```

**Output:**

```
search_results: {
  "results_per_query": {
    "Python web scraping libraries": [...],
    "FastAPI best practices": [...],
    "Redis caching strategies": [...]
  }
}
```

### Example 3: Current News Search

**Input:**

```
search_queries: ["cryptocurrency market news today"]
model: "gpt-4o-mini-search-preview-2025-03-11"
```

**Output:**

```
search_results: {
  "results": [
    {
      "title": "Bitcoin reaches new highs...",
      "url": "https://news.example.com/...",
      "snippet": "Bitcoin trading at record levels as market sentiment turns positive...",
      "published": "2024-11-26"
    }
  ]
}
```

## Common Use Cases

* **Real-time Research**: Get current information on any topic for your workflows
* **News Monitoring**: Track breaking news and latest developments
* **Competitor Analysis**: Search competitor information and market positioning
* **Fact-Checking**: Verify claims against current web information
* **Content Research**: Gather research materials for content generation
* **Market Analysis**: Search for current market trends and data
* **Price Comparison**: Find current prices and product information
* **Event Information**: Get details about upcoming events and announcements

## Error Handling

| Error Type           | Cause                               | Solution                                                                      |
| -------------------- | ----------------------------------- | ----------------------------------------------------------------------------- |
| No Search Queries    | search\_queries is empty or missing | Provide at least one search query                                             |
| Too Many Queries     | More than 10 queries provided       | Reduce to 10 or fewer queries per request                                     |
| Invalid Model        | Model name doesn't exist            | Use gpt-4o-search-preview-2025-03-11 or gpt-4o-mini-search-preview-2025-03-11 |
| Authentication Error | Invalid or missing OpenAI API key   | Verify OpenAI connection is properly configured                               |
| Search Failed        | Web search encountered an error     | Try with simpler queries or check internet connectivity                       |
| Rate Limit Exceeded  | Too many requests in a short time   | Implement delays between requests                                             |
| Empty Results        | No search results found for query   | Refine query terms or try broader searches                                    |
| Timeout Error        | Request took too long to complete   | Try with fewer queries or simpler search terms                                |

## Notes

* **Query Limits**: You can submit 1-10 queries per request. Batch processing multiple queries is more efficient than sequential requests.
* **Model Selection**: Use gpt-4o-search-preview-2025-03-11 for complex queries requiring deep reasoning. Use gpt-4o-mini-search-preview-2025-03-11 for simpler queries to reduce costs.
* **Results Format**: Results may include titles, URLs, snippets, and metadata depending on the search type and model.
* **Current Information**: This node provides real-time web search results, making it ideal for current events and breaking news.
* **Search Quality**: Craft clear, specific search queries for best results. Use quotes for exact phrases and boolean operators if supported.
* **Cost Optimization**: The mini model is significantly cheaper. Use it when full reasoning power isn't needed.
* **Rate Limiting**: OpenAI may rate-limit high volumes of searches. Implement backoff strategies for production use.