Set up External Sources for Genie

External Sources lets you index approved content that lives outside Showpad, such as public websites and Zendesk help centers, so Genie agents can search and cite it alongside your Showpad content.

You configure each source once, Genie crawls it, and the indexed pages become available to the agents you attach it to.

See how it looks

Key features

• Approved knowledge beyond Showpad: index public websites and Zendesk help centers so Genie can draw on them
• Crawl on your terms: set document limits, crawl depth, and automatic recrawling per source
• Cited answers: agents reference external content with a click-through to the original page
• Source-level control: activate, pause, edit, or delete each source, and review exactly which documents were indexed

Note: Genie capabilities depend on your organization's AI Bundle tier.

You need this to succeed

The quick way

1. Go to Admin Settings by clicking the gear icon
2. Scroll to Genie in the left menu and select External Sources
3. Click Create new.
4. Enter a Name, choose the source type, and add the Url.
5. Click Save, then confirm you have permission to crawl the domain.
6. Open the source and click New Crawl.

Do this step by step

Create a source

1. Go to Admin Settings by clicking the gear icon.
2. Scroll to Genie in the left menu and select External Sources.
3. Click Create new.
4. Under Basic settings, enter a Name for the source.
5. Set the Document Limit (the maximum number of documents to index, default 1000).
6. Set the Document Error Limit (how many document errors to allow before the crawl stops, default 50).
7. Check Enable automatic recrawling to keep the source up to date. Recrawls run approximately once per week during off-peak hours.
8. Choose the source type:
   • Web crawls a website. Enter the site's Url, and the crawler follows links from that page.
   • Zendesk crawls a public Zendesk help center. Enter the Url of the Zendesk site.
9. Enter the Url of the site to crawl.

Configure advanced settings (Web sources)

1. Expand Advanced settings.
2. Set the Depth, which is how many links deep the crawler follows from the starting page (default 10).
3. Check Continue even if the page is still requesting data after 30 seconds if pages are slow to load.
4. Keep Block requests to standard ad, consent, telemetry & tracking services enabled to skip non-content requests.
5. Add Additional Blocked Url Patterns to exclude specific paths.
6. Add Starting URLs to give the crawler extra entry points.
7. Add Cookie Presets to apply custom cookies to the browser before the crawl begins.

Save the source

1. Click Save.
2. In the Confirm crawl permission dialog, review the domain.
3. Check I confirm that I have permission to crawl this domain.
4. Click Continue to save.

Run a crawl

1. Open the source from the External Sources list.
2. Make sure the source is set to Active.
3. Click New Crawl.
4. Follow progress on the Crawls tab while the crawl runs.

Review indexed documents

1. Open the Documents tab
2. Use the Filter and Status menus to narrow the list
3. Check the Title and URL of each document. A green check means the document is indexed and available to Genie. A magnifying glass means the page was discovered but not yet indexed

Manage the source

1. Open the Details tab to view the source's settings
2. Use the Active toggle to pause or resume the source
3. Click Edit source to change its settings, or the trash icon to delete it

Note: External Sources is separate from MCP connectors. External Sources index and store web content in Genie's search layer, while connectors give agents direct, action-based access to external systems.

Tip: Attach a source to specific agents in Agent Studio so each agent only draws on the external knowledge that's relevant to it. Suggested labels are: genie, crawl, zendesk, website, knowledge, admin2.