How to Use Web Tools in Openclaw

Table of contents: [Show]

Quick Start

Web tools in Openclaw let your agent search the web and fetch content without writing custom code. This guide shows you how to enable and configure web search and fetch capabilities in under 5 minutes.

For related automation capabilities, check out our guide on Firecrawl Web Scraping for advanced content extraction.

Enable Web Tools

The fastest way to set up web tools is using the interactive configuration wizard:

openclaw configure --section web

This wizard helps you select a search provider and securely stores your API keys in ~/.openclaw/openclaw.json.

Allowlist the Tools

If you use tool allowlists to restrict your agent's capabilities, add these entries to your tools.allow array:

{ "tools": { "allow": ["web_search", "web_fetch", "group:web"] } }

You can also use the shorthand group:web to enable all web-related tools at once.

Working Example

Once configured, your agent can perform web searches directly:

User: Search for the latest Openclaw release notes

Agent: I'll search for that information.
[Tool: web_search - query: "Openclaw latest release notes"]

Results: Openclaw v2.5 released with new channel support...

If you need structured data extraction, you might also want to explore Perplexity Sonar Openclaw for AI-powered search with citations.

Configuration Reference

Your web tools configuration is stored in ~/.openclaw/openclaw.json:

{
  "tools": {
    "web": {
      "searchProvider": "brave",
      "apiKey": "${BRAVE_API_KEY}"
    }
  }
}

Troubleshooting

Tool not found: Verify the tool is in your allowlist and the plugin is enabled
API key errors: Check that your API key is correctly exported or stored in the config file
Rate limits: Most search providers have daily quotas; monitor your usage in the provider dashboard

Best Practices

Use environment variables for API keys instead of hardcoding them
Enable group:web for simpler configuration
Combine web_search with Brave Search Openclaw for privacy-focused results