Skip to main content

Search the web with bulk tools in Sheets

Search the web directly in GPT for Sheets using bulk tools with web search models. These models use fresh web search data, which allows you to generate up-to-date responses about companies, current events, facts, and more, directly in your spreadsheet. You can also fetch content from specific URLs to extract information from web pages.

Choose the right web search model

Choose a web search model that best fits your use case.

Get information that is not in the model's built-in knowledge:

  • Example: Finding the current CEO of a company.

  • How it works:

    All web search models can go beyond their built-in knowledge by accessing information from search engines, like Google or Bing. This allows models, which have compressed historical knowledge, to explore topics in greater depth, with more current data or with a specific focus.

  • Recommended models:

Content fetching

Focus on specific URLs and extract content from them:

  • Example: Extracting the product name and price from the URL of a product page.

  • How it works: Some web search models can access the full content from specific URLs. The full content is added to the context of the models, so they can use it to answer questions about the content or extract information from it.

  • Recommended models:

Supported web search models

Click here for a list of all web search models supported by GPT for Sheets.
ProviderModelWithout API keyWith API key URL fetch
Perplexity logoPerplexitysonarYesYesNo
Perplexity logoPerplexitysonar-proYesYesNo
Perplexity logoPerplexitysonar-reasoning-proNoYesNo
Google logoGooglegemini-2.0-flashNoYesNo
Google logoGooglegemini-2.0-flash-liteNoYesNo
Google logoGooglegemini-2.5-flashYesYesYes
Google logoGooglegemini-2.5-flash-liteNoYesYes
Google logoGooglegemini-2.5-proNoYesYes
Google logoGooglegemini-3-flash-previewNoYesYes
Google logoGooglegemini-3-pro-previewNoYesYes
OpenAI logoOpenAIgpt-4o-mini-search-previewYesYesNo
OpenAI logoOpenAIgpt-4o-search-previewYesYesNo
OpenAI logoOpenAIgpt-5-search-apiYesYesNo

For detailed pricing information, see Model prices.

Use web search models

Use bulk tools to perform web searches with any model that has web search capabilities (🌐).

  1. Open GPT for Sheets.

  2. In the sidebar, select Bulk tools and select the bulk tool you want to use.

    Bulk tools in the sidebar
  3. Open the model switcher.

    Open the model switcher
  4. Enable the web search filter and select a web search model (🌐).

    Filter for and select a web search model
  5. (Optional) Configure the model-specific settings in the sidebar. For Gemini models, make sure web search is enabled.

  6. Set up your bulk tool run. For example, to set up a Custom prompt run:

    Setup Custom
    FieldDescriptionExample
    1

    Prompt to run for each row

    Enter the prompt you want to run for each row.

    Find and summarize reviews for: {'{{Event}}'}.

    3

    Put results in column

    Select the column to put the results in. Cells in these columns won't be overwritten with the results if they contain text.

    B: Information

    4

    Show sources

    Enable this option to write source references in the column to the right of the results column.

    Enabled

  7. Click Run rows.

Your search results appear in the output column. If you enabled source references, they appear in the adjacent column.

Run Custom

For instructions on how to use the tools, see Bulk tools.

Model-specific settings

Use the model-specific settings to configure the web search model for your use case.

SettingDescriptionExample

Web search

Let Gemini decide when a Google search can improve the answer. This allows Gemini to provide more accurate answers and cite verifiable sources beyond its built-in knowledge. Enabling Web search incurs an additional search cost.

Configure Google Gemini web search option

URL context

Gemini 2.5 only

Provide up to 20 URLs (pages, images, PDFs) in your prompt to focus your search. No search cost is incurred, but the content retrieved from URLs counts as input tokens.

Configure Google Gemini URL context option