Loading image...Kiro
  • CLI
  • Powers
  • Autonomous agent
  • Enterprise
  • Pricing
  • Docs
SIGN INDOWNLOADS
Loading image...Kiro
Loading image...Kiro
Product
  • About Kiro
  • CLI
  • Powers
  • Autonomous agent
  • Pricing
  • Downloads
For
  • Enterprise
  • Startups
Resources
  • Documentation
  • Blog
  • Changelog
  • FAQs
  • Report a bug
  • Suggest an idea
  • Billing support
Social
Site TermsLicenseResponsible AI PolicyLegalPrivacy PolicyCookie Preferences
  1. Docs
  2. IDE
  3. Chat
  4. Web tools

Web tools

On this page
  • Limitations

Web access capabilities enable Kiro agent to access current information from the internet in real-time. This feature enables you to get up-to-date answers about topics that may have changed since the model's training data was created. These tools have been designed to not reproduce meaningful chunks of text and it should not be able to access webpages behind paywalls, authentication, and similar access restrictions. Search results may vary over time as internet content changes. Some content may not be accessible through web search due to various restrictions or the nature of the content.

ToolDescription
web_searchTool for searching the web
web_fetchTool for fetching content from a URL
Important

You are responsible for your use of output that incorporates web search or grounded information. You will know when your output includes grounded information from citations or links to the source material. You must retain and display these citations and links in the output if you display the grounded output to a downstream user. If you don't want Kiro to use these tools, create a custom agent that excludes these tools. Citations are provided for output that incorporates web search or grounded information. You can follow a provided citation to the source page.

In the following demo, the agent automatically uses the web_search tool to search the web and return up-to-date results for a query, instead of relying on possibly outdated information in the model's training data. It then uses the web_fetch tool to retrieve the latest contents at a specific URL.

Limitations

  • Size: 10MB maximum per page fetch
  • Timeout: 30 seconds per request
  • Redirects: Maximum 10 redirects followed
  • Content type: Only text/html pages supported
  • Retries: 3 automatic retry attempts on failure
Page updated: December 18, 2025
Subagents
Powers