web

Original：🇺🇸 English

Translated

Web scraping and search via paid API: crawl any URL with cascade escalation (fast HTTP -> browser -> stealth+proxy), bulk crawl multiple URLs, and search the web via Exa. Uses x_payment tool for automatic USDC micropayments ($0.005/crawl, $0.01/search). Use as escalation when built-in web_fetch fails or is blocked. Use when: (1) web_fetch returned empty/blocked content, (2) scraping JS-rendered or anti-bot protected pages, (3) bulk-crawling multiple URLs, (4) searching the web by query via Exa.

2installs

Sourcecascade-protocol/agentbox

Added on2026-03-11

NPX Install

npx skill4agent add cascade-protocol/agentbox web

SKILL.md Content

View Translation Comparison →

Web Scraping & Search

Paid web crawling and search API at

https://web.surf.cascade.fyi

. Crawl costs $0.005 USDC, search costs $0.01 USDC per call via x402 on Solana. Use the

x_payment

tool for all requests. Use this when the built-in

web_fetch

tool fails or is blocked (anti-bot, JS-rendered pages, paywalled content).

Endpoints

Crawl Web Pages

Fetch and extract content from any URL. Automatically escalates through fetching tiers if blocked: fast HTTP -> headless browser -> stealth browser with proxy.

x_payment({
  "url": "https://web.surf.cascade.fyi/v1/crawl",
  "method": "POST",
  "params": "{\"url\": \"https://example.com\", \"format\": \"markdown\"}"
})

Body parameters:

Param	Type	Default	Description
url	string	-	Single URL to crawl (required if no `urls` )
urls	string[]	-	Multiple URLs to bulk crawl (required if no `url` )
format	string	markdown	Output format: `markdown` , `html` , or `text`
selector	string	-	CSS selector to extract specific elements
proxy	boolean	false	Force proxy usage (auto-enabled on stealth tier)

Single URL response:

json

{
  "status": 200,
  "content": ["# Page Title\n\nPage content in markdown..."],
  "url": "https://example.com/"
}

Bulk URL response:

json

[
  {"status": 200, "content": ["..."], "url": "https://example.com/"},
  {"status": 200, "content": ["..."], "url": "https://other.com/"}
]

Cascade escalation: The crawler automatically tries increasingly powerful methods:

get - Fast HTTP request (cheapest, handles most static sites)
fetch - Headless Chromium (for JS-rendered pages)
stealthy_fetch - Stealth browser with proxy (bypasses Cloudflare, anti-bot)

Escalation triggers: HTTP 403/429/503, empty content, or content under 100 characters.

Search the Web

Search the web using Exa's search API. Returns titles, URLs, and text snippets.

x_payment({
  "url": "https://web.surf.cascade.fyi/v1/search",
  "method": "POST",
  "params": "{\"query\": \"x402 protocol crypto payments\", \"num_results\": 10}"
})

Body parameters:

Param	Type	Default	Description
query	string	required	Search query
num_results	1-20	5	Number of results to return

Response:

json

{
  "results": [
    {
      "title": "Page Title",
      "url": "https://example.com/page",
      "snippet": "Relevant text excerpt from the page..."
    }
  ]
}

Usage Patterns

Scrape a page as markdown

x_payment({
  "url": "https://web.surf.cascade.fyi/v1/crawl",
  "method": "POST",
  "params": "{\"url\": \"https://docs.solana.com\", \"format\": \"markdown\"}"
})

Extract specific elements with a CSS selector

x_payment({
  "url": "https://web.surf.cascade.fyi/v1/crawl",
  "method": "POST",
  "params": "{\"url\": \"https://news.ycombinator.com\", \"format\": \"html\", \"selector\": \".titleline\"}"
})

Bulk crawl multiple URLs

x_payment({
  "url": "https://web.surf.cascade.fyi/v1/crawl",
  "method": "POST",
  "params": "{\"urls\": [\"https://example.com\", \"https://httpbin.org/get\"], \"format\": \"text\"}"
})

Search and then crawl top results

First search:

x_payment({
  "url": "https://web.surf.cascade.fyi/v1/search",
  "method": "POST",
  "params": "{\"query\": \"ERC-8004 agent identity standard\", \"num_results\": 5}"
})

Then crawl the most relevant result:

x_payment({
  "url": "https://web.surf.cascade.fyi/v1/crawl",
  "method": "POST",
  "params": "{\"url\": \"https://eips.ethereum.org/EIPS/eip-8004\", \"format\": \"markdown\"}"
})

Cost

Crawl: $0.005 USDC per call (single or bulk)
Search: $0.01 USDC per call

All payments on Solana mainnet. Each request is a separate call. For bulk crawls, one payment covers all URLs in the batch.

Errors

HTTP	Meaning
400	Invalid parameters (check body format)
402	Payment required (handled automatically by x_payment)
429	Too many concurrent requests (retry later)
502	Upstream crawl/search error

Web Scraping & Search

Paid web crawling and search API at

https://web.surf.cascade.fyi

. Crawl costs $0.005 USDC, search costs $0.01 USDC per call via x402 on Solana. Use the

x_payment

tool for all requests. Use this when the built-in

web_fetch

tool fails or is blocked (anti-bot, JS-rendered pages, paywalled content).

Endpoints

Crawl Web Pages

Fetch and extract content from any URL. Automatically escalates through fetching tiers if blocked: fast HTTP -> headless browser -> stealth browser with proxy.

x_payment({
  "url": "https://web.surf.cascade.fyi/v1/crawl",
  "method": "POST",
  "params": "{\"url\": \"https://example.com\", \"format\": \"markdown\"}"
})

Body parameters:

Param	Type	Default	Description
url	string	-	Single URL to crawl (required if no `urls` )
urls	string[]	-	Multiple URLs to bulk crawl (required if no `url` )
format	string	markdown	Output format: `markdown` , `html` , or `text`
selector	string	-	CSS selector to extract specific elements
proxy	boolean	false	Force proxy usage (auto-enabled on stealth tier)

Single URL response:

json

{
  "status": 200,
  "content": ["# Page Title\n\nPage content in markdown..."],
  "url": "https://example.com/"
}

Bulk URL response:

json

[
  {"status": 200, "content": ["..."], "url": "https://example.com/"},
  {"status": 200, "content": ["..."], "url": "https://other.com/"}
]

Cascade escalation: The crawler automatically tries increasingly powerful methods:

get - Fast HTTP request (cheapest, handles most static sites)
fetch - Headless Chromium (for JS-rendered pages)
stealthy_fetch - Stealth browser with proxy (bypasses Cloudflare, anti-bot)

Escalation triggers: HTTP 403/429/503, empty content, or content under 100 characters.

Search the Web

Search the web using Exa's search API. Returns titles, URLs, and text snippets.

x_payment({
  "url": "https://web.surf.cascade.fyi/v1/search",
  "method": "POST",
  "params": "{\"query\": \"x402 protocol crypto payments\", \"num_results\": 10}"
})

Body parameters:

Param	Type	Default	Description
query	string	required	Search query
num_results	1-20	5	Number of results to return

Response:

json

{
  "results": [
    {
      "title": "Page Title",
      "url": "https://example.com/page",
      "snippet": "Relevant text excerpt from the page..."
    }
  ]
}

Usage Patterns

Scrape a page as markdown

x_payment({
  "url": "https://web.surf.cascade.fyi/v1/crawl",
  "method": "POST",
  "params": "{\"url\": \"https://docs.solana.com\", \"format\": \"markdown\"}"
})

Extract specific elements with a CSS selector

x_payment({
  "url": "https://web.surf.cascade.fyi/v1/crawl",
  "method": "POST",
  "params": "{\"url\": \"https://news.ycombinator.com\", \"format\": \"html\", \"selector\": \".titleline\"}"
})

Bulk crawl multiple URLs

x_payment({
  "url": "https://web.surf.cascade.fyi/v1/crawl",
  "method": "POST",
  "params": "{\"urls\": [\"https://example.com\", \"https://httpbin.org/get\"], \"format\": \"text\"}"
})

Search and then crawl top results

First search:

x_payment({
  "url": "https://web.surf.cascade.fyi/v1/search",
  "method": "POST",
  "params": "{\"query\": \"ERC-8004 agent identity standard\", \"num_results\": 5}"
})

Then crawl the most relevant result:

x_payment({
  "url": "https://web.surf.cascade.fyi/v1/crawl",
  "method": "POST",
  "params": "{\"url\": \"https://eips.ethereum.org/EIPS/eip-8004\", \"format\": \"markdown\"}"
})

Cost

Crawl: $0.005 USDC per call (single or bulk)
Search: $0.01 USDC per call

All payments on Solana mainnet. Each request is a separate call. For bulk crawls, one payment covers all URLs in the batch.

Errors

HTTP	Meaning
400	Invalid parameters (check body format)
402	Payment required (handled automatically by x_payment)
429	Too many concurrent requests (retry later)
502	Upstream crawl/search error

web

NPX Install

Tags

SKILL.md Content

Web Scraping & Search

Endpoints

Crawl Web Pages

Search the Web

Usage Patterns

Scrape a page as markdown

Extract specific elements with a CSS selector

Bulk crawl multiple URLs

Search and then crawl top results

Cost

Errors

web

NPX Install

Tags

SKILL.md Content

Web Scraping & Search

Endpoints

Crawl Web Pages

Search the Web

Usage Patterns

Scrape a page as markdown

Extract specific elements with a CSS selector

Bulk crawl multiple URLs

Search and then crawl top results

Cost

Errors