exa

name	exa
description	Neural and semantic web search using Exa via StableEnrich (enrichx402). Use when the user needs intelligent search, retrieving page contents, or getting AI-generated answers with citations.

Exa

Search the web using Exa's neural search engine through StableEnrich's x402-enabled API. Exa excels at semantic/meaning-based search, extracting page contents, and generating answers with citations.

When to Use

• User needs semantic or neural search (finding pages by meaning, not just keywords)
• User needs to extract clean text content from specific URLs
• User wants an AI-generated answer backed by web search citations
• User asks for research papers, company pages, news, or tweets by category

Common Operations

Search

• URL: https://stableenrich.dev/api/exa/search
• Method: POST
• Pricing: ~$0.005 per request

Request:

obulx -X POST -H "Content-Type: application/json" \
  -d '{
    "query": "latest research on LLM reasoning",
    "type": "auto",
    "numResults": 5,
    "contents": {
      "text": true
    }
  }' \
  "https://stableenrich.dev/api/exa/search"

Parameters:

Parameter	Type	Default	Description
query	string	required	Search query
type	string	"auto"	"neural", "auto", "keyword"
numResults	integer	10	Number of results (max 100)
category	string	—	"company", "research paper", "news", "tweet", "personal site", "financial report", "people"
includeDomains	array	—	Restrict results to these domains
excludeDomains	array	—	Exclude these domains
startPublishedDate	string	—	ISO 8601 date filter start
endPublishedDate	string	—	ISO 8601 date filter end
contents.text	bool/object	—	Include full text. Object form: {"maxCharacters": 2000}
contents.highlights	bool/object	—	Include relevant snippets. Object form: {"maxCharacters": 500, "query": "specific focus"}
contents.summary	object	—	LLM-generated summary. {"query": "summarize the key findings"}

Response:

{
  "requestId": "abc123",
  "resolvedSearchType": "neural",
  "results": [
    {
      "id": "https://example.com/article",
      "url": "https://example.com/article",
      "title": "Article Title",
      "author": "Author Name",
      "text": "Full page content...",
      "highlights": ["Relevant snippet 1", "Relevant snippet 2"],
      "summary": "LLM-generated summary of the page"
    }
  ],
  "costDollars": { "total": 0.005 }
}

Find Similar

• URL: https://stableenrich.dev/api/exa/findSimilar
• Method: POST
• Pricing: ~$0.005 per request

Request:

obulx -X POST -H "Content-Type: application/json" \
  -d '{
    "url": "https://example.com/interesting-article",
    "numResults": 5,
    "contents": {
      "text": true
    }
  }' \
  "https://stableenrich.dev/api/exa/findSimilar"

Response: Same format as Search with semantically similar pages.

Get Contents

• URL: https://stableenrich.dev/api/exa/contents
• Method: POST
• Pricing: ~$0.001 per request

Request:

obulx -X POST -H "Content-Type: application/json" \
  -d '{
    "urls": ["https://example.com/page1", "https://example.com/page2"],
    "text": true
  }' \
  "https://stableenrich.dev/api/exa/contents"

Parameters:

Parameter	Type	Default	Description
urls	array	required	URLs to extract content from
text	bool/object	—	Include full text. Object form: {"maxCharacters": 5000}
highlights	bool/object	—	Include relevant snippets
summary	object	—	LLM-generated summary
maxAgeHours	integer	—	Cache freshness; triggers live crawl if stale
subpages	integer	—	Number of subpages to also crawl

Response:

{
  "results": [
    {
      "url": "https://example.com/page1",
      "title": "Page Title",
      "author": "Author Name",
      "text": "Full extracted page content..."
    }
  ],
  "costDollars": { "total": 0.001 }
}

Answer

• URL: https://stableenrich.dev/api/exa/answer
• Method: POST
• Pricing: ~$0.01 per request

Request:

obulx -X POST -H "Content-Type: application/json" \
  -d '{
    "query": "What are the latest developments in x402 payments?",
    "text": true
  }' \
  "https://stableenrich.dev/api/exa/answer"

Response: JSON object with an AI-generated answer and an array of cited source results.

Notes

• Exa's neural search is best for semantic queries where meaning matters more than exact keyword matching
• Use "type": "keyword" for exact-match searches, "neural" for semantic, "auto" to let Exa decide
• The category parameter significantly improves result quality when you know the content type
• contents is useful when you already have URLs and just need to extract their text
• All endpoints route through StableEnrich's x402-enabled API at stableenrich.dev
• Settlement is in USDC on Base chain, handled automatically by the Obul proxy