Add a website source
Use this guide to save a public web page into an Atlas project so it can be searched, cited, and used in chat.
Before you start
Website import works best on pages where the main content is visible without logging in. Avoid pages that require a login, a paywall, or a consent screen to access the content you want.
Steps
- Copy the page URL from your browser's address bar.
- Open the Atlas project where the source belongs.
- Open the add-source panel and paste the URL.
- Wait for Atlas to fetch and process the page.
- Open the new source and check that the correct content was captured.
Choosing the right URL
Use the direct URL of the specific article, report, or page, not the site homepage or a search results page.
Good URLs to import:
- A specific article or essay
- A documentation page with clear body text
- A public report or reference page
URLs to avoid:
- Site homepages with mostly navigation
- Search results pages
- Pages behind a login or paywall
- Infinite scroll feeds
- Cookie or consent screens
Check the captured content
Open the source and compare it with the original page. Confirm that the title, main body text, and key headings look correct.
If the source shows only navigation text or very little content, try a more specific URL or a print-friendly version of the page.
Result
After the source processes, you can ask questions about its content, receive citations linking back to specific passages, and mention the source in notes.
Troubleshooting
| Symptom | Likely cause |
|---|---|
| Import fails | The page may be private, blocked, or unsupported |
| Captured content is thin | The page renders content dynamically or has a paywall |
| Wrong text appears | The URL redirects to a landing page or consent screen |