FIREHOSE v5

Smart Web Extraction API for AI Agents & LLMs

Recursive crawling, JS rendering, intelligent extraction, structured data.
Beats Firecrawl at 1/1000th the price. No signup required.

Try it FREE View Pricing

Features

Smart Extraction

Trafilatura-powered article extraction. Strips nav, ads, footers. Returns clean text + metadata (title, author, date, language).

JS Rendering

Headless Chromium in an isolated jail. Renders SPAs, React, Vue. Auto-fallback when static extraction yields too little content.

Recursive Crawl

BFS crawl with parallel workers. Sitemap discovery, robots.txt compliant. Up to 200 pages per request.

Structured Data

Extract JSON-LD and microdata natively. No LLM needed. Schema.org, OpenGraph, product data — all included.

Batch Processing

Send up to 10 URLs in one request. Parallel fetching with 4 workers. One response, all results.

Privacy-First

No logs, no tracking. Free tier needs no API key. Requests proxied through isolated FreeBSD jails.

Endpoints

/v1/markdown

Convert a single URL or raw HTML to Markdown.

# Smart extraction with metadata
curl -s "https://synthetic-context.net/v1/markdown?url=https://example.com"

# With JSON-LD + microdata extraction
curl -s "https://synthetic-context.net/v1/markdown?url=https://example.com&extract=jsonld,microdata"

# JS rendering (paid tier)
curl -s -H "X-API-Key: sk_YOUR_KEY" \
  "https://synthetic-context.net/v1/markdown?url=https://spa-app.com&js=1"

# Batch POST
curl -s -X POST https://synthetic-context.net/v1/markdown \
  -H "Content-Type: application/json" \
  -d '{"urls": ["https://a.com", "https://b.com"], "extract": "jsonld"}'

/v1/crawl

Recursively crawl an entire website. Discovers pages via sitemap + link extraction.

# Crawl up to 10 pages, depth 2
curl -s -X POST https://synthetic-context.net/v1/crawl \
  -H "Content-Type: application/json" \
  -d '{"url": "https://docs.example.com", "max_pages": 10, "max_depth": 2}'

# Crawl with JS rendering + structured data (paid tier)
curl -s -X POST https://synthetic-context.net/v1/crawl \
  -H "Content-Type: application/json" \
  -H "X-API-Key: sk_YOUR_KEY" \
  -d '{"url": "https://spa.com", "max_pages": 50, "js": true, "extract": "jsonld"}'

Response Format

{
  "markdown": "# Article Title\n\nClean extracted text...",
  "meta": {"title": "...", "author": "...", "date": "...", "language": "en"},
  "jsonld": [{"@type": "Article", "name": "..."}],
  "microdata": [{"@type": "https://schema.org/Product", "name": "..."}],
  "source": "https://example.com",
  "length": 4521
}

Try It

Enter a URL and see the result:

Pricing

	Free	Paid	Premium
Price	0	5 sats/req	Contact
Requests/day	100	10,000	Unlimited
JS Rendering	-	Yes	Yes
Crawl max pages	5	50	200
API Key	Not needed	X-API-Key header	X-API-Key header
Batch URLs	10	10	10
Structured data	Yes	Yes	Yes

vs Competitors

Feature	Firehose	Firecrawl	Crawl4AI	Jina Reader
Smart extraction	Yes	Yes	Yes	Yes
JS rendering	Yes	Yes	Yes	Yes
Recursive crawl	Yes	Yes	Yes	No
Sitemap discovery	Yes	Yes	Yes	No
JSON-LD extraction	Native (0 LLM)	Via LLM ($)	Via LLM ($)	No
Microdata extraction	Native	No	No	No
robots.txt	Yes	Yes	Yes	No
No signup	Yes	No	Yes	No
Price	5 sats (~$0.005)	From $38/mo	Self-hosted	From $0.01/req