clawd
8cc0dcb167
- Moved 4 skills: browser-testing, claude-multimedia, exa-search, gravl-research - Moved 14 agents: architect, backend-dev, browser-tester, coach, data, flight, frontend-dev, gravl-pm, gravl-researcher, nutritionist, research, reviewer, staging, update - Created symlinks from ~/clawd/skills and ~/clawd/agents back to hub - Single source of truth in claude-agents-skills repo
79 lines
2.5 KiB
Markdown
79 lines
2.5 KiB
Markdown
---
|
|
name: exa-search
|
|
description: Web search, code search, company research, and deep research via Exa AI. Use for real-time web queries, finding code examples, researching companies, finding people, or starting deep research tasks. Triggers on web search needs, code lookups, company/people research.
|
|
---
|
|
|
|
# Exa Search
|
|
|
|
Exa AI search via their hosted MCP server. Provides web search, code context, company research, people search, and deep research capabilities.
|
|
|
|
## Quick Start
|
|
|
|
```bash
|
|
# Web search
|
|
~/clawd/skills/exa-search/scripts/exa-cli.mjs search "latest AI agent frameworks 2026"
|
|
|
|
# Code search
|
|
~/clawd/skills/exa-search/scripts/exa-cli.mjs code "Python async retry with exponential backoff"
|
|
|
|
# Company research
|
|
~/clawd/skills/exa-search/scripts/exa-cli.mjs company "Anthropic AI"
|
|
|
|
# People search
|
|
~/clawd/skills/exa-search/scripts/exa-cli.mjs people "CTO machine learning startups Stockholm"
|
|
|
|
# Crawl specific URL
|
|
~/clawd/skills/exa-search/scripts/exa-cli.mjs crawl "https://example.com/article"
|
|
|
|
# Deep research (async)
|
|
~/clawd/skills/exa-search/scripts/exa-cli.mjs research "State of quantum computing 2026"
|
|
~/clawd/skills/exa-search/scripts/exa-cli.mjs research-check <task-id>
|
|
```
|
|
|
|
## Available Tools
|
|
|
|
| Tool | Use Case |
|
|
|------|----------|
|
|
| `search` | General web search, news, current events |
|
|
| `search-advanced` | Filtered search (domains, dates, content type) |
|
|
| `code` | Code examples, docs, GitHub, StackOverflow |
|
|
| `crawl` | Full content from specific URL |
|
|
| `company` | Company info, news, competitors |
|
|
| `people` | LinkedIn profiles, professional bios |
|
|
| `research` | Start deep AI research task |
|
|
| `research-check` | Check research task status |
|
|
|
|
## API Key (Optional)
|
|
|
|
Free tier has rate limits. Add your key for higher limits:
|
|
|
|
```bash
|
|
export EXA_API_KEY="your-key-here"
|
|
```
|
|
|
|
Get key: https://dashboard.exa.ai/api-keys
|
|
|
|
## Advanced Search Options
|
|
|
|
```bash
|
|
# Filter by domain
|
|
exa-cli.mjs search-advanced "AI news" --domains techcrunch.com,wired.com
|
|
|
|
# Date range
|
|
exa-cli.mjs search-advanced "OpenAI announcements" --after 2026-01-01
|
|
|
|
# More results
|
|
exa-cli.mjs search "topic" --num 20
|
|
```
|
|
|
|
## Tool Details
|
|
|
|
See [references/tools.md](references/tools.md) for full parameter documentation.
|
|
|
|
## Tips
|
|
|
|
- **Code search**: Include language name ("Python", "Go", "TypeScript") for better results
|
|
- **Company research**: Start with company name, then drill into specifics
|
|
- **Deep research**: Returns task ID; poll with `research-check` until complete
|
|
- **Rate limits**: Free tier ~1 req/sec; add API key for more
|