# Exa Fetch

> Read a webpage's full content using Exa. Gets clean markdown text from any URL. Use when you have a specific URL and need its contents. Supports batching multiple URLs. Use exa-web-search instead when you need to find pages first.

- Skill: `exa-labs/exa-fetch` (Agent Skill)
- Install (CLI): `npx skillmds@latest add exa-labs/exa-fetch`
- Raw SKILL.md: https://api.skillmd.com/api/skills/exa-labs/exa-fetch/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Docs & Writing
- Author: exa-labs (https://skillmd.com/u/exa-labs)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/exa-labs/exa-fetch

---


# Fetch URL (Exa)

Extract content from: $ARGUMENTS

## Tool

Use `web_fetch_exa` to get the full text of a webpage.

## When to Use

- You have a specific URL and need its contents
- Reading documentation pages, blog posts, or articles
- Extracting data from a known webpage
- Following up on a search result to get full content

Trigger phrases: "read this page", "fetch this URL", "extract content from", "get the page at", "what does this link say", or when a user shares a URL.

Use `web_search_exa` instead when you don't have a URL and need to find pages first.

## How to Fetch

`web_fetch_exa` accepts:
- `urls` (array of strings, required) — the URLs to extract content from
- `maxCharacters` (number, optional) — max characters to return per page (default: 3000)

### Character Limits

- Quick summary → `maxCharacters: 1000-2000`
- Standard read → `maxCharacters: 3000` (default)
- Full page content → `maxCharacters: 5000-10000`

### Batching

You can fetch multiple URLs in a single call:

```
web_fetch_exa({
  "urls": ["https://example.com/page1", "https://example.com/page2"],
  "maxCharacters": 3000
})
```

## Output Format

1. Present the key information from the page
2. If the user asked a specific question, answer it using the page content
3. Include the source URL

## Before You Start

Check if Exa MCP is installed by confirming `web_fetch_exa` is available in your tool list.

## If the MCP Server Is Not Connected

### You (the AI) must:

1. **Stop immediately** — do NOT try to fetch the URL yourself, do NOT use websearch, and do NOT try to answer from your own knowledge.
2. Inform the user that Exa MCP is not connected.

### Request the user to:

1. Run the `/exa-setup` command in Cursor to add the Exa MCP server.
2. Go to **Cursor Settings → Tools & MCP** and make sure Exa is toggled **on** (green).
3. Retry their fetch after completing the above steps.