Search Web Node
The Search Web node performs internet searches using Google's Custom Search API to retrieve current information from the web. It returns structured results containing URLs, titles, and content snippets, with support for up to 100 results per query. Optional domain restrictions limit searches to specific websites or trusted sources.
How It Works
When the node executes, it constructs a search request with query terms and configuration, sends it to Google's search service, and retrieves matching web pages with their URLs, titles, and content snippets. The search results are structured data objects that can be processed by downstream nodes for information extraction, content analysis, or response generation in AI workflows.
The node requires Google API credentials (API key and Custom Search Engine ID) to authenticate with Google's search service. Credentials can be provided directly in the node configuration or referenced from workflow variables for security and reusability. Domain restrictions limit results to specific websites, useful for searching within trusted sources or organizational knowledge bases. The node handles pagination automatically when requesting multiple results, fetching up to 100 results in a single execution.
Search results are returned as a structured response object containing an array of result items, where each item includes the page URL, title, snippet (preview text), and additional metadata. Variable interpolation across all parameters enables dynamic search queries based on user questions, conversation context, or data from previous workflow steps.
Configuration Parameters
Output Field
Output Field (Text, Required): Workflow variable where search results are stored.
The output is a web search response object containing results (array of search results), query (original search query), and total_results (estimated total matches). Each result item contains url (page URL), title (page title), snippet (preview text), displayLink (display domain), and additional metadata.
Common naming patterns: web_search_results, search_results, google_results, web_data.
Search Query
Search Query (Text, Optional): Search terms to query the web.
Concise queries (typically 1-6 words) yield best results, as Google's search algorithm is optimized for natural language queries. Variable interpolation with ${search_query} supports dynamic queries from user input or conversation context. If not provided in the node configuration, the query must be available from workflow context or LLM-generated parameters.
Number of Results
Number of Results (Number, Default: 10): Number of search results to return (1-100).
Google Custom Search API has a maximum limit of 100 results per request. Higher values may take longer due to pagination, as Google returns results in batches of 10. Lower values (10-20) provide faster responses and are sufficient for most use cases. API quota limits should be considered when setting this value.
Google API Key
Google API Key (Text, Required): Google API key for authenticating with Custom Search API.
Variable interpolation with ${google_api_key} supports secure credential management. Storing API keys in workflow variables rather than hardcoding improves security and enables credential rotation. API keys can be obtained from the Google Cloud Console by enabling the Custom Search API.
Google CSE ID
Google CSE ID (Text, Required): Google Custom Search Engine ID.
The CSE ID determines search behavior, including which sites to search, SafeSearch settings, and other search preferences configured in the Custom Search Engine. Variable interpolation with ${google_cse_id} is supported. Custom Search Engines can be created at https://programmablesearchengine.google.com/.
Restrict to Sites
Restrict to Sites (Array, Optional): Domain names to limit search results to specific websites.
Example: ['example.com', 'example.org'] restricts results to only those domains. Leave empty to search the entire web. Domain restrictions are applied in addition to any site restrictions configured in the Custom Search Engine settings.
Common Parameters
This node supports common parameters shared across workflow nodes, including Stream Output Response, Streaming Messages, and Logging Mode. For detailed information, see Common Parameters.
Best Practices
- Concise queries with 1-6 keywords yield more precise results than full sentences
- Variable interpolation passes user questions or extracted keywords from conversation context for dynamic search
- Storing Google API credentials in workflow variables improves security, enables credential rotation, and allows credential sharing across workflows
- Set Number of Results based on use case: 10-20 for quick top results, 50-100 for comprehensive research, considering API quota consumption
- Restrict to Sites filters results to trusted domains, improving relevance and reducing downstream filtering
- Monitor Google Custom Search API usage; the free tier has 100 queries per day, with paid tiers for higher volumes
Limitations
- API quota limits: Google Custom Search API has daily quota limits (100 queries per day for free tier). Exceeding quotas causes search failures until the quota resets or the plan is upgraded.
- Result limit: Maximum 100 results per search request. Larger result sets require multiple searches with different queries.
- API key required: The node requires valid Google API credentials (API key and CSE ID). Without credentials, searches fail with authentication errors.
- Search freshness: Search results reflect Google's index, which may not include the most recent web content. Newly published content may take time to appear.
- Content access: The node returns search result metadata (URLs, titles, snippets) but does not fetch actual web page content. Additional nodes are needed to retrieve full page content.
- Rate limiting: Google API enforces rate limits on requests per second. Rapid consecutive searches may be throttled or rejected.