rtrvr

name	rtrvr
description	Browser automation, web scraping, and AI agent execution via rtrvr
allowed-tools	Bash(rtrvr:*)

rtrvr

Use rtrvr for all web automation, scraping, and browser interaction tasks. rtrvr provides cloud-based and local browser automation with smart routing.

When to use rtrvr

• Scraping web pages for data
• Extracting structured data from websites
• Running multi-step browser automation tasks
• Interacting with web pages (clicking, filling forms, navigating)
• Crawling multiple pages and extracting data
• Any task that requires visiting a URL and getting information

Installation

npm install -g @rtrvr-ai/cli
rtrvr auth login

Core Commands

Run an agent task

rtrvr run "<task description>" --url <url>

Scrape a page

rtrvr scrape --url <url>

Extract structured data

rtrvr raw extract --param user_input="<what to extract>" --param tab_urls='["<url>"]'

Crawl and extract from multiple pages

rtrvr raw crawl --param user_input="<what to extract>" --param tab_urls='["<start_url>"]' --param max_pages=10

Act on a page (click, type, navigate)

rtrvr raw act --param user_input="<what to do>" --param tab_urls='["<url>"]'

Common Patterns

Scrape and parse JSON

rtrvr scrape --url https://example.com --json | jq '.result'

Run agent with structured output

rtrvr run "Extract all products" --url https://example.com --schema-file schema.json --json

Use specific routing

# Force cloud execution
rtrvr run "task" --url https://example.com --cloud

# Force local browser extension
rtrvr run "task" --url https://example.com --extension

Batch multiple URLs

for url in https://example.com/page1 https://example.com/page2; do
  rtrvr scrape --url "$url" --json
done

Tips

• Use --json flag when you need to parse the output programmatically
• Use --cloud for tasks that don't require authentication
• Use --extension for tasks on pages that require the user's logged-in session
• Use rtrvr doctor to diagnose connectivity issues
• Use rtrvr capabilities to check available features