Web fetch

Web fetch reads a web page and returns its text. On Cue OS, Cue uses it to read a source it found or one you gave it. Cue CLI lets you fetch a page yourself when you already have its address.

What you can do

• Read the main text of a public page, with its title.
• Use a page as source material for a summary or another task.
• Get the result as JSON for a script.

On the web

Cue OS has no separate page for this. Cue's hosted agent can read web pages, so you can paste a link in a chat and ask Cue to summarize it.

On iPhone and Android

The iPhone and Android apps are coming; store review is pending. You can use Cue OS in your phone browser now.

With Cue CLI

1. Set up Cue CLI.

2. Fetch a page:

   cue web fetch "https://example.com"

3. Read the header lines: the address, the page title, how the page was fetched, and how much text came back. The page text follows.

4. Add --json when a script needs the result:

   cue web fetch "https://example.com" --json

Limits today

• The request goes straight from your computer to the page. It does not sign in, so pages behind a login, a paywall, or heavy scripts may come back empty or partial.
• Choosing another fetch service with --provider does not take effect in the current build.
• Treat fetched text as material to read, not as instructions for your agent.

Related

Cue CLI on cueos.ai · Web search · Search and everyday tools · cue web reference