Exa

Neural web search with page contents, similar-page discovery and cited answers.

4 actions API key 4 read0 write Calls api.exa.ai

Connect

TypeScript

const conn = await setu.connections.initiate({ userId: "user_1", toolkit: "exa" });
console.log(conn.connectUrl); // show this to your user
await conn.waitUntilActive();

Permissions requested

This toolkit uses a single credential and requests no scopes.

Actions

Get a cited answerreadexa.answerAsk a question and get an answer written from live web search, with the source citations.

Input	Type		Description
`query`	string	Required	The question to answer
`text`	boolean	Optional	Include the full text of cited pages
`model`	string	Optional	Answer model
`systemPrompt`	string	Optional	Extra instructions for the answer

TypeScript

await setu.execute({
  userId: "user_1",
  action: "exa.answer",
  input: {
    "query": "<query>"
  },
});

Find similar pagesreadexa.find_similarFind web pages similar to a given URL, optionally excluding the source domain or limiting to certain domains.

Input	Type		Description
`url`	string	Required	The URL to find similar pages for
`numResults`	integer	Optional	Number of results, up to 100 (default 10)
`includeDomains`	array of string	Optional	Only return results from these domains
`excludeDomains`	array of string	Optional	Never return these domains
`excludeSourceDomain`	boolean	Optional	Exclude results from the same domain as the url
`contents`	object	Optional	Options for page contents, e.g. {"highlights": true}

TypeScript

await setu.execute({
  userId: "user_1",
  action: "exa.find_similar",
  input: {
    "url": "<url>"
  },
});

Get page contentsreadexa.get_contentsFetch the text, highlights or summary of specific pages by URL (or Exa document id), up to 100 at a time.

Input	Type		Description
`urls`	array of string	Required	URLs to fetch, 1 to 100
`text`	boolean	Optional	Return the full page text
`highlights`	boolean	Optional	Return relevant highlight snippets
`maxAgeHours`	integer	Optional	Maximum cache age in hours; 0 always fetches fresh, -1 always uses cache

TypeScript

await setu.execute({
  userId: "user_1",
  action: "exa.get_contents",
  input: {
    "urls": []
  },
});

Search the webreadexa.searchSearch the web with Exa and get ranked results with titles, URLs and optional page text or highlights.

Input	Type		Description
`query`	string	Required	The search query
`type`	string	Optional	Search type (default auto)
`numResults`	integer	Optional	Number of results, 1 to 100 (default 10)
`category`	string	Optional	Focus on a category
`includeDomains`	array of string	Optional	Only return results from these domains
`excludeDomains`	array of string	Optional	Never return these domains
`startPublishedDate`	string	Optional	Only results published after this ISO 8601 date-time
`endPublishedDate`	string	Optional	Only results published before this ISO 8601 date-time
`contents`	object	Optional	Options for page contents, e.g. {"highlights": true} or {"text": true}

TypeScript

await setu.execute({
  userId: "user_1",
  action: "exa.search",
  input: {
    "query": "<query>"
  },
});