exa-search

name	Exa Search
description	Advanced web search with precise date filtering and content type selection. Use when you need academic papers, GitHub repositories, research content, or specific date ranges. Handles neural search (semantic understanding), keyword search, and content type filtering (research papers, GitHub, news, PDFs). Perfect for research, competitive analysis, and content discovery.
metadata	{"requires":{"env":["CLAW0X_API_KEY"]}}

Exa Search

Cloud skill by Claw0x — powered by Claw0x Gateway API.

Advanced web search with precise date filtering, content type selection, and neural search. Perfect for research, competitive analysis, and specialized content discovery.

Requires Claw0x API key. Sign up at claw0x.com to get your key.

Prerequisites

1. Sign up at claw0x.com
2. Create API key in Dashboard
3. Set environment variable:

   # Add to ~/.openclaw/.env
   CLAW0X_API_KEY=ck_live_...
   

Pricing

$0.005 per successful call. Failed calls are free.

• Pay only for successful responses (2xx status)
• No monthly fees, no subscriptions
• Get started with $5 free credit

Quick Reference

When This Happens	Do This	What You Get
Need academic papers from specific dates	Use category: "research paper" + date range	Filtered research results
Find GitHub projects from 2024	Use category: "github" + start_published_date: "2024-01-01"	Recent open-source projects
Semantic search for concepts	Use search_type: "neural"	Intent-based results
Exact keyword matching	Use search_type: "keyword"	Traditional search results

How It Works — Under the Hood

Exa is a next-generation search engine built specifically for AI agents and researchers. Unlike traditional search engines, Exa offers:

1. Neural Search (Semantic Understanding)

Exa's neural search understands query intent, not just keywords:

• Query: "papers about attention mechanisms improving transformers"
• Traditional search: matches keywords "attention", "mechanisms", "transformers"
• Exa neural: understands you want research on how attention improves transformer architecture

2. Precise Date Filtering

Unlike Tavily's coarse-grained time ranges (day/week/month/year), Exa supports exact dates:

• Tavily: time_range: "month" (last 30 days)
• Exa: start_published_date: "2024-03-15", end_published_date: "2024-03-22" (exact 7-day window)

3. Content Type Filtering

Exa can restrict results to specific content types:

• Research papers — academic publications, arXiv, journals
• GitHub — open-source repositories and code
• News — news articles and press releases
• PDFs — PDF documents
• Company sites — corporate websites
• Personal sites — blogs and personal pages
• Tweets — Twitter/X posts

4. Autoprompt Optimization

Exa can automatically optimize your query for better results:

• Input: "transformer papers"
• Optimized: "research papers on transformer architecture improvements"

Exa vs Tavily: When to Use Which?

Feature	Tavily	Exa	Winner
Date filtering	time_range (coarse)	start/end date (precise)	🏆 Exa
Search mode	basic/advanced (depth)	neural/keyword (algorithm)	🏆 Exa
Content types	general/news	7+ types (papers, GitHub, PDF)	🏆 Exa
AI answer	✅ Built-in	❌ Not available	🏆 Tavily
Speed	1-2s (basic), 3-5s (advanced)	2-3s	🏆 Tavily
Best for	Quick lookups, news, general info	Research, analysis, specialized content	—

Use Tavily when: You need a quick answer, general web search, or news lookup.

Use Exa when: You need precise date ranges, specific content types (papers, GitHub), or semantic search.

Real-World Use Cases

Scenario 1: Academic Research

Problem: Agent needs to find transformer architecture papers published in Q1 2024.

Tavily limitation:

{ query: "transformer architecture", time_range: "year" }
// Returns papers from last 365 days (too broad)

Exa solution:

{
  query: "transformer architecture improvements",
  category: "research paper",
  start_published_date: "2024-01-01",
  end_published_date: "2024-03-31",
  search_type: "neural"
}
// Returns only Q1 2024 research papers

Scenario 2: GitHub Project Discovery

Problem: Find Rust web frameworks created in 2024.

Exa solution:

{
  query: "rust web framework",
  category: "github",
  start_published_date: "2024-01-01",
  num_results: 10
}

Scenario 3: Competitive Analysis

Problem: Find companies similar to a competitor.

Exa solution:

{
  query: "AI agent platforms",
  category: "company",
  exclude_domains: ["competitor.com"],
  num_results: 10
}

Scenario 4: News Tracking (Precise Dates)

Problem: Find AI regulation news from March 15-22, 2024.

Tavily limitation:

{ query: "AI regulation", time_range: "week" }
// Returns last 7 days from today (not specific week)

Exa solution:

{
  query: "AI regulation",
  category: "news",
  start_published_date: "2024-03-15",
  end_published_date: "2024-03-22"
}
// Returns exact date range

Input Schema

Field	Type	Required	Default	Description
query	string	yes	—	Search query (1–500 chars)
search_type	string	no	"neural"	"neural", "keyword", or "auto"
start_published_date	string	no	—	ISO 8601 date: "2024-01-01"
end_published_date	string	no	—	ISO 8601 date: "2024-12-31"
category	string	no	—	Content type (see below)
num_results	number	no	5	Number of results (1–10)
include_domains	string[]	no	—	Only include these domains
exclude_domains	string[]	no	—	Exclude these domains
use_autoprompt	boolean	no	true	Auto-optimize query
include_text	boolean	no	false	Include full page text
text_length_limit	number	no	—	Max text length (chars)
highlights	boolean	no	false	Highlight key sentences

Content Categories

• "company" — Company websites
• "research paper" — Academic papers
• "news" — News articles
• "github" — GitHub repositories
• "tweet" — Twitter/X posts
• "pdf" — PDF documents
• "personal site" — Blogs and personal pages

Search Types

• "neural" (default) — Semantic search, understands intent
• "keyword" — Traditional keyword matching
• "auto" — Automatically choose best mode

Output Schema

{
  results: Array<{
    title: string;              // Page title
    url: string;                // Page URL
    published_date: string | null;  // Publication date
    author: string | null;      // Author (if available)
    score: number;              // Relevance score (0-1)
    text?: string;              // Full text (if requested)
    highlights?: string[];      // Key sentences (if requested)
    summary?: string;           // Summary (if available)
    category: string | null;    // Content type
    domain: string;             // Domain name
  }>;
  autoprompt_string?: string;   // Optimized query
  result_count: number;         // Number of results
  _meta: {
    skill: "exa-search";
    latency_ms: number;
    search_type: string;
    date_filtered: boolean;
  };
}

Integration Recipes

OpenClaw Agent

import { Claw0xClient } from '@claw0x/sdk';

const claw0x = new Claw0xClient(process.env.CLAW0X_API_KEY);

// Research papers from specific date range
const papers = await claw0x.call('exa-search', {
  query: 'large language model reasoning',
  category: 'research paper',
  start_published_date: '2024-01-01',
  end_published_date: '2024-03-31',
  search_type: 'neural',
  num_results: 10
});

console.log(`Found ${papers.result_count} papers`);
papers.results.forEach(paper => {
  console.log(`${paper.title} (${paper.published_date})`);
  console.log(paper.url);
});

LangChain Agent (Python)

from claw0x import Claw0xClient

client = Claw0xClient(api_key=os.environ['CLAW0X_API_KEY'])

# Find GitHub projects
repos = client.call('exa-search', {
    'query': 'rust web framework',
    'category': 'github',
    'start_published_date': '2024-01-01',
    'num_results': 5
})

for repo in repos['results']:
    print(f"{repo['title']}: {repo['url']}")

Custom Agent (JavaScript)

const response = await fetch('https://api.claw0x.com/v1/call', {
  method: 'POST',
  headers: {
    'Authorization': `Bearer ${process.env.CLAW0X_API_KEY}`,
    'Content-Type': 'application/json'
  },
  body: JSON.stringify({
    skill: 'exa-search',
    input: {
      query: 'AI regulation news',
      category: 'news',
      start_published_date: '2024-03-15',
      end_published_date: '2024-03-22'
    }
  })
});

const data = await response.json();
console.log(data.results);

Error Handling

Error	Cause	Solution
CLAW0X_API_KEY is required	Missing API key	Set environment variable
Invalid input	Wrong input format	Check input schema
search_type must be one of: neural, keyword, auto	Invalid search type	Use valid enum value
category must be one of: ...	Invalid category	Use valid content type
API error (500)	Server error	Retry or contact support

Why Use Via Claw0x?

• Unified billing: One API key for all skills
• Atomic pricing: Pay per call, not per month
• Zero cost on failure: Failed calls don't charge
• Production-ready: 99.9% uptime, <100ms latency
• Security scanned: OSV.dev integration
• No Exa API key needed: Claw0x handles upstream authentication

About Claw0x

Claw0x is the native skills layer for AI agents — providing unified API access, atomic billing, and quality control.

Explore more skills: claw0x.com/skills

GitHub: github.com/kennyzir/exa-search