How to walk through links in Chrome

To walk through links in Chrome with Tab Automation, choose a listing page, use For Each Link to collect matching http(s) URLs, and add the steps that should run on each URL.

The automation uses one assigned tab. It opens a link there, completes the inner steps, then opens the next link in that same tab. It does not create a new tab for every URL.

Collect the links and add the loop

You’ll need Tab Automation installed. Custom automations are free; Explorer unlocks the full Dashboard. Google Search is coming soon. The side-panel tour covers first-time setup.

1. Open the page that holds the links — a pull-request list, a queue, a search results page.
2. Open the Tab Automation side panel.
3. Click New Automation. Under Tab & Startup, turn off Always start this automation in a new tab, then pick that tab.
4. Click Add Step and choose For Each Link.
5. Click Pick links, then click one of the repeated links on the page. The picker saves a rule for finding those links again, plus the URLs available when you pick. Picking also writes that page as the Start URL.
6. Add the steps that should run inside Each Link. For example, add Wait / Delay before interacting with a page, If / Else to check it, or Send Alert in a branch that should notify or POST a webhook.
7. Optional Limit how many → Up To caps the included list. Open the links accordion to Ignore URLs you do not want visited; ignored URLs stay visible in the review list and stay skipped on later cycles, but consume none of the limit. Skip visited defaults to Timed, which remembers visits for 24 hours. Always keeps skipping remembered URLs; Never allows revisits.
8. Turn Repeat off for one pass, or configure it to repeat the graph. Choose Save & Run.

Each Link starts empty. Steps already on the main chain stay above the new loop. Add a Wait inside Each Link when each destination needs time to load; the loop does not insert one for you.

With 100 collected links, ignoring links 1–10 and setting Up To to 5 selects links 11–15. It selects fewer if fewer included URLs remain. Skip visited is applied afterward and does not replace skipped URLs with later ones. Pasted lists have no Up To control.

Fallback order for page links and pasted lists

For On page and Paste list, starting a run returns to Start and tries the sources in this order:

1. Live re-scrape. It applies the saved selector to Start and uses the matching http(s) URLs found now.
2. Paste list. If the live re-scrape is empty, valid URLs entered through Paste list take priority.
3. Saved collection. If neither source has URLs, it uses the matching URLs saved when you clicked Pick links.
4. Error. If all three sources are empty, Start reports an error.

The automation does not scrape whichever tab happens to be active. Start remains the listing page.

Use Paste list when you already know the destinations or when the source page cannot expose them with one selector. Enter one http(s) URL per line.

Collect from an XML sitemap

Choose Sitemap for a public UTF-8 XML page sitemap (<urlset>) or sitemap index (<sitemapindex>), uncompressed or gzip-compressed. Enter its URL and click Collect links to preview pages from nested indexes within the collection limits. Hostname and path filters narrow the list. Filters and Ignore apply before Up To, which selects the first included URLs in order. Starting a run and each repeat cycle fetch the sitemap again.

Sitemap sources do not use the page/paste fallback order above. A failed child sitemap stops collection without running a partial list or using another source. HTML sitemaps, plain-text files, RSS/Atom feeds, and arbitrary XML are not supported. See supported sitemap formats, controls, and limits.

What happens to the assigned tab

Each Link includes a Go To This Link card with an Open action beside each included collected URL. Use it to visit a sample in the assigned tab and configure inner steps there. Back to Start returns to the saved listing page. Opening a sample preserves Start and creates no results or visited-link records; the card itself is not an executable step.

For example, collect product names and prices into CSV with Save Text or save product images from multiple pages with Save Images. Both guides show how to check selections on another product before running the list.

After navigation, the automation runs the inner steps. When they finish, it opens the next URL. For Each Link adds no separate delay between links, so an inner Wait / Delay is the place to give each destination time.

Steps can also sit before or after the loop. The End block can stop after one pass or use Repeat to collect and walk again after its wait. Repeat can also stop after a duration.

This is different from auto refreshing multiple tabs, where every tab remains separate and has its own timer.

Check progress and failures

The automation row shows progress such as 2 of 8 and a status for the current action. Open Activity to see collection, target, step, and cycle outcomes.

If a selector misses, confirm that Start is still the intended listing page, then use Pick links again. If destinations fail because they load slowly, lengthen the inner Wait / Delay.

Use Pause to stop the run. Resume continues from saved progress; an interrupted step may need to run again. For a paused input or storage failure, follow the error’s retry instructions before resuming. Restart begins a new run and collects the list again, but remembered visits still apply. Clear Skip visited history or choose Never when you deliberately want to revisit those pages.

If you only need the listing page itself to reload, use Refresh Page plus Repeat instead. The privacy policy explains local storage and page permissions.