Exa Search

Exa is a neural search API optimized for finding specific types of content - especially people, companies, and structured information that traditional search struggles with.

When to Use Exa vs. web_search

Use Exa when:

• Finding people (LinkedIn profiles, personal sites)
• Finding companies (company pages, profiles)
• Searching within specific content types (research papers, PDFs, tweets, GitHub repos)
• Building prospect/contact lists
• Need semantic/embeddings-based search
• Want structured extraction (summaries with JSON schema)

Use web_search (Brave) when:

• General web search
• Recent news/events
• Simple lookups

Setup

Get API key from: https://dashboard.exa.ai/api-keys

Store in environment:

export EXA_API_KEY="your_api_key_here"

Core Features

1. Category Search

Search within specific content types using category:

• people - LinkedIn profiles (high quality)
• company - Company pages (high quality)
• research paper - Academic papers
• news - News articles
• pdf - PDF documents
• github - GitHub repositories
• tweet - Twitter/X posts
• personal site - Personal websites/blogs
• financial report - Financial documents

Example - Find people:

curl -X POST https://api.exa.ai/search \
  -H "x-api-key: $EXA_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "query": "medical billing managers at hospitals in Texas",
    "category": "people",
    "numResults": 20
  }'

Example - Find companies:

curl -X POST https://api.exa.ai/search \
  -H "x-api-key: $EXA_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "query": "revenue cycle management software companies",
    "category": "company",
    "numResults": 10
  }'

2. Search Types

• auto (default) - Intelligently picks best method
• neural - Embeddings-based semantic search
• deep - Query expansion + comprehensive results (supports additionalQueries)
• fast - Streamlined/faster version

Example - Deep search with query variations:

curl -X POST https://api.exa.ai/search \
  -H "x-api-key: $EXA_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "query": "prior authorization software",
    "type": "deep",
    "additionalQueries": [
      "pre-authorization automation",
      "prior auth platforms"
    ],
    "numResults": 20
  }'

3. Content Extraction

Get full text, highlights, or summaries:

curl -X POST https://api.exa.ai/search \
  -H "x-api-key: $EXA_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "query": "latest LLM benchmarks",
    "category": "research paper",
    "numResults": 5,
    "contents": {
      "text": {
        "maxCharacters": 2000
      },
      "highlights": {
        "numSentences": 3,
        "highlightsPerUrl": 2
      },
      "summary": {
        "query": "key findings and results"
      }
    }
  }'

Structured summaries with JSON schema:

curl -X POST https://api.exa.ai/search \
  -H "x-api-key: $EXA_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "query": "Series B funding announcements healthcare",
    "category": "news",
    "numResults": 10,
    "contents": {
      "summary": {
        "schema": {
          "type": "object",
          "properties": {
            "companyName": {"type": "string"},
            "fundingAmount": {"type": "string"},
            "leadInvestor": {"type": "string"},
            "date": {"type": "string"}
          },
          "required": ["companyName", "fundingAmount"]
        }
      }
    }
  }'

4. Date Filtering

Filter by published date or crawl date:

curl -X POST https://api.exa.ai/search \
  -H "x-api-key: $EXA_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "query": "Medicare coverage changes",
    "category": "news",
    "startPublishedDate": "2025-01-01T00:00:00.000Z",
    "numResults": 10
  }'

5. Domain Filtering

Include or exclude specific domains:

curl -X POST https://api.exa.ai/search \
  -H "x-api-key: $EXA_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "query": "machine learning papers",
    "includeDomains": ["arxiv.org", "openreview.net"],
    "numResults": 20
  }'

Common Use Cases

Building Prospect Lists

Find 100 medical billing managers:

curl -X POST https://api.exa.ai/search \
  -H "x-api-key: $EXA_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "query": "medical billing manager at hospital",
    "category": "people",
    "numResults": 100,
    "contents": {
      "text": {
        "maxCharacters": 500
      }
    }
  }'

Process results to extract names, titles, companies, LinkedIn URLs.

Finding Companies in a Space

Find RCM software companies:

curl -X POST https://api.exa.ai/search \
  -H "x-api-key: $EXA_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "query": "revenue cycle management software company",
    "category": "company",
    "numResults": 50,
    "contents": {
      "summary": {
        "query": "What does this company do and who are their customers?"
      }
    }
  }'

Research/Competitive Analysis

Find recent competitor blog posts:

curl -X POST https://api.exa.ai/search \
  -H "x-api-key: $EXA_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "query": "prior authorization automation",
    "includeDomains": ["competitor1.com", "competitor2.com"],
    "startPublishedDate": "2024-11-01T00:00:00.000Z",
    "numResults": 20,
    "contents": {
      "text": true
    }
  }'

Finding Recent Papers/News

Latest AI research on specific topic:

curl -X POST https://api.exa.ai/search \
  -H "x-api-key: $EXA_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "query": "LLM agent frameworks",
    "category": "research paper",
    "startPublishedDate": "2025-01-01T00:00:00.000Z",
    "numResults": 10,
    "contents": {
      "highlights": {
        "numSentences": 2,
        "highlightsPerUrl": 3
      }
    }
  }'

Important Constraints

Category-Specific Limitations

For people and company categories, these filters are NOT supported:

• startPublishedDate / endPublishedDate
• startCrawlDate / endCrawlDate
• includeText / excludeText
• excludeDomains

For people category:

• includeDomains only accepts LinkedIn domains

Using unsupported parameters returns 400 error.

Result Limits

• neural / auto: max 100 results
• deep: max 100 results
• Contact sales for higher limits

Tips

1. Be specific in queries - "medical billing manager at hospital in Texas" beats "billing manager"
2. Use deep search for comprehensive results when you need variations
3. Combine with date filters for recent content
4. Request structured summaries when building lists/databases
5. Use highlights over full text for cost efficiency when you need snippets
6. Category search is powerful - use it! Much better results than generic search

Cost-Efficient Patterns

• Use highlights instead of full text when possible (cheaper)
• Set maxCharacters limits on text extraction
• Use numResults wisely - start small, increase if needed
• auto search type usually best - only use deep when necessary

Environment Variable

# Add to your shell profile or .env
export EXA_API_KEY="your_api_key_here"

More Info

• Dashboard: https://dashboard.exa.ai
• API Docs: https://exa.ai/docs
• Get API Key: https://dashboard.exa.ai/api-keys