Seed URLs: Usage and Examples

What is a Seed URL?

A Seed URL is a starting point you give to the AI agent to begin its investigation. While the AI usually gathers information through open web searches, setting a Seed URL lets the system start from that specific URL first.

Typical Use Cases and Examples

1. Monitoring Competitor Press Releases

By setting a competitor's "News" or "Press Release" page URL, you can reliably capture their latest moves without missing anything.

• Example: https://competitor-example.com/news/ or https://prtimes.jp/main/html/searchrlp/company_id/XXXXX

2. Routine Observation of Specific Blogs or Media

Specify an industry tech blog or specialized media outlet that puts out the latest trends. This is effective for reducing noise in the collected data.

• Example: https://techcrunch.com/category/artificial-intelligence/

3. Tracking Official Guidelines and Government Announcements

Target pages of government offices where legal revisions and new guidelines are announced so you can rapidly generate reports on compliance changes.

• Example: https://www.fsa.go.jp/news/

Seed URLs alone do not define what the AI should extract from those pages. If you also want stable output criteria, pair them with How to Write Effective Research Instructions so the source list and the research question stay aligned.

Important Notes

• A Seed URL is currently limited to one URL per job. If you need to watch multiple sites, split them into separate jobs or add constraints in the instruction.
• Depending on the site's structure (e.g., heavy dynamic rendering via JavaScript, or strong anti-bot protections), the system may not be able to retrieve information successfully. In such cases, please consider switching to standard keyword searches.