<aside>

On this page:

</aside>

<aside>

Expanse Docs

🧠 Product Docs

• Expanse 101
• Key features

🤖 Prompting Tips

LLM Directory

Prompting Tips & Tricks

Prompt Library

💻 Use Cases

For Work

For Fun

📚 Resources

Prompt Engineering Resources

🍄 About Expanse

Meet the team

</aside>

Getting Web content into Expanse

When working with Expanse, having quality content to feed your AI is essential. While Expanse doesn't directly connect to the internet, there's a clever way to bring web content into your conversations with minimal effort.

<aside>

TLDR: Add https://r.jina.ai/ to the beginning of the URL you want to scrape, then copy the results into Expanse.

</aside>

Why raw web copying doesn't work well

Web pages are designed for human eyes, not AI processing. When you simply copy and paste from a website, you're bringing along a lot of unnecessary elements that can confuse large language models.

Elements that confused LLMs:

• Navigation menus and sidebars
• Advertisements and popups
• Complex formatting that doesn't translate well

These elements can cause the AI to focus on irrelevant parts or misinterpret the structure of your content, leading to less accurate responses.

Converting to markdown: a better approach

Markdown provides a clean, structured format that works beautifully with LLMs like those in Expanse. It is a clearer for LLMs to understand and more widely accepted in accepted in most applications.

Benefits of markdown conversion:

• Clean structure - Removes visual clutter while preserving organization
• Semantic preservation - Maintains headings, lists, and other important elements
• Better AI comprehension - Helps the AI understand the content hierarchy
• Human readability - Still easy for you to read and edit if needed

<aside> 💡

Converting to markdown essentially gives you the best of both worlds—content that's optimized for AI processing while remaining accessible to human readers.

</aside>

How to get clean Web Content for Expanse

The simplest method is using Jina.ai's reader service:

1. Find the webpage you want to extract content from
2. Add https://r.jina.ai/ to the beginning of the URL
3. Visit this new combined URL
4. Copy the clean markdown version that appears

For example:

• Original article: https://defector.com/this-is-the-last-thing-you-see-before-you-die-if-you-are-a-crab
• Converted URL: https://r.jina.ai/<https://defector.com/this-is-the-last-thing-you-see-before-you-die-if-you-are-a-crab>

Example: fetch up-to-date code docs

Why this matters

The quality of input you provide directly affects the quality of output you receive. By using this method to scrape web content, you::

1. Save time compared to manual reformatting
2. Improve the accuracy of AI responses
3. Create a more seamless workflow between web research and AI assistance

This approach helps you get the most out of Expanse by ensuring your conversations are based on clean, well-structured information rather than messy web formatting.

<aside> ⚠️

Note: Always respect copyright and terms of service when scraping web content. This technique should be used for personal research or with content you have permission to use.

</aside>

Quick alternative: use a search model to fetch the context for you

Didn’t find what you were looking for?