Make AI Web Search

the ai web search app is currently in beta there may be changes to functionality or pricing ai web search is a built in app that lets you find and retrieve up to date information from the internet directly within your it searches across websites, news articles, and other online sources to provide relevant, structured results — perfect for enriching workflows with fresh data, fact checking, or content discovery all searches run securely within , without depending on external third party integrations the app retrieves and analyzes information from static html web pages by reading the initial page source it acts as a simple html fetcher and does not function as a real browser as a result, it cannot execute javascript, render client side content, or interact with dynamically generated pages such as single page applications (spas) it also cannot reliably extract data from sites where content loads dynamically, appears in canvases or svgs, or is heavily generated by client side scripts key benefits no external accounts no need to connect or manage a separate account everything works within your existing environment privacy focused your data is never stored or used for ai training instant setup start using the app immediately in your platform there is no need to create any connection credit usage the module uses gpt 5 nano and the credit usage is based on the amount of input and output tokens consumed in each run input tokens depend on the prompt you enter, and output tokens are based on the response after each run, you can see exactly how many tokens were used in the module's output the module uses gpt 5 nano and the credit usage is based on the amount of input and output tokens consumed in each run input tokens depend on the prompt you enter, and output tokens are based on the response after each run, you can see exactly how many tokens were used in the module's output the table below shows how input and output tokens convert to make credits for openai models the table below shows how input and output tokens convert to make credits for openai models model name model name input tokens input tokens output tokens output tokens gpt 5 nano gpt 5 nano 18080 tokens per credit 18080 tokens per credit 2260 tokens per credit 2260 tokens per credit for more detailed information about credits, refer to the https //help make com/credits documentation ai web search modules you can use the following module to build your the current features and functionality are still in beta and are actively being improved as such, pricing is subject to change and will be updated as the app evolves generate a response generates a response from the input provided field description text enter your search query for the ai you can include details on how you want the response to be formatted parse json in response enable to parse the ai’s output as json if parsing fails, the raw text response will be returned city enter a city to refine the search location for example, san francisco country select or map a country to prioritize search results relevant to that country content from the selected country will appear higher in web search results to map a country manually, enter its two letter iso code for example, us region enter a region to narrow the search location for example, california timezone select or map a timezone to adjust results based on local time for example, america/los angeles example retrieve information from a website in this use case, you’ll use the make ai web search app to retrieve information from a website you’ll find the most recent posts on the make blog, collect their urls and publication dates, and extract their content the scenario returns all data in a structured json format step 1 set up your first make ai web search module first, you need to add the make ai web search app to your and configure it to retrieve one of today’s blog post urls add a make ai web search > generate a response module to your in the text field, enter the prompt to retrieve the information you need example prompt get one of today's blog post urls nest these items under the urls get the date of the last blog post formatted yyyy mm dd, blog title output in json wrapped by \[startjson] and \[endjson] do not ask for clarification and do it now https //www make com/en/blog switch on the advanced settings option in the parse json in response field, select yes to format the response as json step 2 add an iterator click save step 2 add an iterator next, add an iterator module to your to handle each blog post result as a separate item add the iterator module to your in the array field, map the result from the previous make ai web search module c lick save step 3 extract the blog post content finally, add another make ai web search module to extract the content of the blog post add a make ai web search > generate a response module to your after the iterator in the text field, enter the prompt to extract the content from the url example prompt extract the blog post body after entering the prompt, map the url value from the iterator module to the text field to extract the retrieved content click save you’ve now set up the to retrieve and extract content from a website, with the data processed step by step for each url and returned in json format templates you can look for ai web search templates in https //www make com/en/templates , where you'll find thousands of pre created