# Exa Search

> Real-time web search using Exa AI. Use when you need current information, fact-checking, researching unfamiliar topics, or finding documentation. Triggers on requests mentioning search, lookup, find online, current information, or research.

- Skill: `majiayu000/exa-search` (Agent Skill, multi-file: 2 files)
- Install (CLI): `npx skillmds@latest add majiayu000/exa-search`
- Raw SKILL.md: https://api.skillmd.com/api/skills/majiayu000/exa-search/raw
- Safety review: pending (external: skill-scanner PASS, skillspector PASS)
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Docs & Writing, AI & ML
- Author: majiayu000 (https://skillmd.com/u/majiayu000)
- Updated: 2026-09-09
- Page: https://skillmd.com/skills/majiayu000/exa-search

---


# Exa Web Search

Search the web for current information using the Exa AI search engine.

## When to Use

- Need information beyond training data cutoff
- Fact-checking claims or statistics
- Researching unfamiliar topics
- Finding official documentation or sources
- Getting current news or recent developments

## Quick Start

Use the `mcp__exa__web_search` tool:

```python
# Basic search
result = mcp__exa__web_search(query="TypeScript 5.0 new features")

# With more results
result = mcp__exa__web_search(query="MCP protocol specification", numResults=10)
```

## Parameters

| Parameter | Type | Required | Default | Description |
|-----------|------|----------|---------|-------------|
| `query` | string | Yes | - | Search query |
| `numResults` | number | No | 5 | Number of results to return |

## Response Format

Returns search results with:
- `title` - Page title
- `url` - Source URL
- `snippet` - Text excerpt
- `content` - Full page content (when available)

## Examples

### Research a topic
```
Search for: "latest React 19 features and changes"
```

### Find documentation
```
Search for: "Anthropic Claude API authentication docs"
```

### Fact check
```
Search for: "current TypeScript version 2025"
```

## Best Practices

1. **Be specific** - Include relevant keywords and context
2. **Use quotes** - For exact phrases: `"error message text"`
3. **Add context** - Include technology/domain: `Python asyncio tutorial`
4. **Limit results** - Use 3-5 for quick lookups, 10+ for research