⚡ Built for LangChain, CrewAI & AutoGPT agents

One API call. Any URL. Structured JSON back.

Shell City routes across Firecrawl, ScrapingBee, and Browserless automatically. Your agent never hits a rate limit. $0.001/page. Free: 1,000 pages/month. No credit card.

Try it now

# Scrape any URL → clean markdown
curl -X POST https://shellcity.polsia.app/v1/scrape \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"url": "https://example.com"}'

# Response: structured JSON, clean markdown
{
  "success": true,
  "markdown": "# Example Domain\n\nThis domain...",
  "credits_used": 1
}

// Drop into any LangChain / CrewAI agent
const res = await fetch('https://shellcity.polsia.app/v1/scrape', {
  method: 'POST',
  headers: {
    'Authorization': `Bearer ${process.env.SHELLCITY_API_KEY}`,
    'Content-Type': 'application/json'
  },
  body: JSON.stringify({ url: 'https://example.com' })
});
const { markdown } = await res.json();
// → "# Example Domain\n\nThis domain..."

# Works with CrewAI, AutoGPT, any Python agent
import requests

res = requests.post(
    "https://shellcity.polsia.app/v1/scrape",
    headers={"Authorization": f"Bearer {api_key}"},
    json={"url": "https://example.com"}
)
markdown = res.json()["markdown"]
# → clean markdown, ready for your LLM

Free: 1,000 pages. No credit card required.

Works with 🦜 LangChain 🚢 CrewAI 🤖 AutoGPT ⚡ AutoGen any HTTP client

Routes across 3+ providers 99.9% uptime SLA $0.001/page Free: 1,000 pages

How it works

Three steps. Your agent is unblocked.

Enter your email, get a key

No contracts, no credit card, no sales calls. Your API key and 1,000 free pages arrive instantly.

Call the scrape endpoint

One POST request. Pass a URL, get clean structured JSON back. Shell City handles JavaScript rendering, CAPTCHAs, and provider routing automatically.

Pay only for what you use

$0.001 per page. No monthly minimums. Your agent's cost scales linearly with what it delivers — nothing leaks to idle subscriptions.

Pricing

Free to start. Pennies at scale.

No subscriptions. No seat fees. Top up when you need to — unused credits don't expire.

Free tier

1,000 pages included

✓ Full API access
✓ No credit card
✓ Instant key delivery

Pay-per-use

$0.001

per page · no minimum

✓ JS rendering included
✓ Auto CAPTCHA handling
✓ Provider failover

Extraction

$0.005

per call · structured JSON

✓ Custom JSON schema
✓ Products, contacts, tables
✓ Zero parsing code

Why Shell City

Every capability an AI agent runs into a wall without.

🌐

Web Scraping

Scrape any URL. Get clean markdown or structured JSON. Handles JavaScript rendering, anti-bot bypasses, and proxies automatically.

From $0.001/page

🔍

Web Search

Real-time search results without maintaining a search API subscription. Google, Bing, or news — one unified interface.

From $0.003/call

🤖

Browser Automation

Full headless browser sessions. Click, type, navigate, screenshot. Your agent controls a real browser without managing infrastructure.

From $0.01/session

📊

Data Extraction

Structured data from any page. Product listings, contact info, pricing tables — extracted and normalized automatically.

From $0.005/call

🔒

CAPTCHA Solving

Agents don't wait for humans. Solve reCAPTCHA, hCaptcha, and Cloudflare challenges programmatically.

From $0.002/solve

📤

Email Verification

Validate email addresses before your agent sends. Catch bounces, disposables, and invalid mailboxes in real time.

From $0.001/check

LangChain Integration

Drop into any LangChain agent

Two tools, one install. ShellCityScrapeTool and ShellCityExtractTool — built for the LangChain agent loop.

Install

npm install shellcity-langchain

Quick start

const { createShellCityTools } = require('shellcity-langchain');
const { createReactAgent } = require('langchain/agents');
const { ChatOpenAI } = require('@langchain/openai');

// One line to get both tools
const { tools } = createShellCityTools(process.env.SHELLCITY_API_KEY);

const agent = createReactAgent({
  llm: new ChatOpenAI({ model: 'gpt-4o-mini' }),
  tools,  // scrape + extract, ready to go
});

View on npm →