# How to let Claude.ai read any URL

Published: 2025-12-04T21:00:00.000Z
Updated: 2026-02-03T00:00:00.000Z
Canonical: https://wow.pjh.is/journal/claude-ai-access-any-url

> **Update 2026-02-03:** Claude.ai can now read the text of any web page URL. The instructions below are redundant.

Claude Code can, with your permission, read any web page.

The Claude.ai web app, by contrast, has a security sandbox that strictly limits the domains it can access.

To work around this, use the [Claude Zapier Connector](https://www.claude.com/connectors/zapier) with the [URL to Text](https://urltotext.com/) service:

![Claude crawling the AI Wow blog](https://wow.pjh.is/static/images/blog/claude-crawl-any-url.jpeg)

I won't explain the steps in detail here. If you can't easily figure this out for yourself, then you shouldn't enable this workaround.[^1]

The one tip I *will* give: for the sake of speed, specify default settings in your [system instructions](https://claude.ai/settings/general). Something like this:

**Prompt**

```

# TOOL USE

## Zapier URL to Text defaults

When using the Zapier `url_to_text_get_text_of_url` tool, always use these settings unless I specify otherwise:
* `url`: the URL I provide
* `output_format`: "markdown"
* `extract_main_content`: "false"
* `render_javascript`: "false"
* `instructions`: "" (empty string)

Use this tool only when I ask you to fetch URL content via Zapier.

```

[^1]: The default URL restrictions are there to protect you from [prompt injection](https://en.wikipedia.org/wiki/Prompt_injection). You should never let Claude crawl URLs that you do not trust.