Web Search & URL Access

ModelMonkey can access the internet to help you research topics, gather information, and read web pages. This is useful when you need external data, want to fact-check information, or need to reference online documentation.

There are two main capabilities:

Web Search
Search the web using natural language queries. ModelMonkey returns relevant results with titles, descriptions, and links that you can explore further.

URL Access
Open and read the content of specific web pages. ModelMonkey converts web content to readable text and can search within long pages to find relevant sections.

When you ask ModelMonkey to search for information, it queries the web and returns up to 10 relevant results by default. You can request more results (up to 20) if needed.

What You Get
Each search result includes:

• Title: The page title
• URL: The link to the page
• Description: A snippet describing the page content
• Page Age: How recently the page was updated (when available)

Example Requests

• "Search for average salary data by industry in the US"
• "Find tutorials on Excel VLOOKUP alternatives"
• "Look up the current exchange rate for USD to EUR"
• "Search for 5 competitors of [company name]"

ModelMonkey will display the search results and can use them to inform your spreadsheet work, such as adding reference data or validating existing information.

ModelMonkey can open and read the content of any publicly accessible web page. This is useful when you need to:

• Extract specific data from a webpage
• Understand documentation or reference material
• Pull information from online articles or reports
• Verify facts against authoritative sources

How It Works
When you ask ModelMonkey to open a URL, it:

1. Fetches the page content
2. Converts it to readable text (removing ads, navigation, etc.)
3. Breaks it into searchable chunks
4. Uses semantic search to find the most relevant sections based on your query

Example Requests

• "Open https://example.com/report and find the revenue figures"
• "Read this documentation page and explain how authentication works"
• "Go to [URL] and extract the pricing information"
• "Check this webpage and tell me what the main conclusions are"

When a web page is long, ModelMonkey automatically finds the most relevant sections based on your query instead of reading the entire page.

Reading More Content
After finding relevant content, you can ask ModelMonkey to read more:

• "Show me more of that section"
• "What comes before this?"
• "Continue reading from where you left off"

ModelMonkey will navigate through the page to give you additional context as needed.

Cached Content
Web pages are cached temporarily, so follow-up questions about the same URL are faster. If you need the latest version of a page that may have changed, mention that you want fresh content.

Market Research
Search for competitor information, industry trends, or market data to inform your analysis.

"Search for the top 10 SaaS companies by revenue and add them to my spreadsheet"

Data Validation
Verify spreadsheet data against authoritative online sources.

"Check if these company headquarters locations are correct by searching online"

Reference Data
Look up conversion rates, standards, or reference values.

"Search for ISO country codes and add them to column C"

Documentation Lookup
Read technical documentation or guides while working on your spreadsheet.

"Open the Google Sheets API documentation and find the rate limits"

News and Updates
Find recent news or updates relevant to your data.

"Search for recent news about [company name] from the last month"

Access Restrictions

• Pages requiring login or authentication cannot be accessed
• Some websites block automated access
• Paywalled content is not accessible
• Very dynamic pages (heavy JavaScript) may not render completely

Content Handling

• Images and videos are not processed—only text content
• Tables in web pages may not preserve their exact formatting
• PDFs linked from web pages require the File Uploads feature

Search Limitations

• Maximum of 20 search results per query
• Search results reflect publicly indexed content
• Very recent content may not appear in search results immediately

Caching

• Web page content is cached temporarily for faster follow-up queries
• Recent changes to web pages may not be immediately reflected

Be Specific with Search Queries
Instead of "search for company data," try "search for Apple Inc annual revenue 2023." More specific queries yield more relevant results.

Provide Context When Opening URLs
When opening a URL, describe what you're looking for: "Open [URL] and find the pricing section" works better than just "Open [URL]."

Combine with Spreadsheet Operations
Web research is most powerful when combined with spreadsheet actions:

• "Search for competitor pricing and add it to a new column"
• "Look up the CEO names for these companies and fill in column D"

Verify Important Data
For critical data, consider checking multiple sources:

• "Search for this statistic and find at least 2 sources to verify"

Use Search Results as Starting Points
Search results give you an overview—if you need details from a specific result, ask ModelMonkey to open that URL and read the full content.