Thirdwatch thirdwatch.dev

What do you need?

Choose the fastest path to useful data. You can switch approaches at any time.

What Thirdwatch does

Thirdwatch maintains a catalog of production-grade web scrapers on the Apify platform and exposes them through a single Model Context Protocol (MCP) server. LLM-driven applications — Claude, Cursor, ChatGPT custom GPTs, internal agents — can use Thirdwatch as a structured-data layer for the public web.

Every scraper returns a documented field set. Every result is billed pay-per-event, with FREE / BRONZE / SILVER / GOLD volume tiers. There's no subscription, no minimum, no commitment. 300+ users across 10+ countries currently use the catalog; 1M+ records have been served.

We specialize in keeping difficult public-data sources useful after their APIs, markup, or anti-bot systems change. For high-volume requirements, missing fields, or a broken workflow, contact support@thirdwatch.dev. Feature requests from real users take priority over speculative catalog growth.

What's in the catalog

82 live scrapers across 16 categories:

How we built it

Each scraper picks the cheapest reliable extraction path: pure HTTP where the site cooperates, TLS-fingerprint spoofing with impit when it doesn't, and a stealth browser (Camoufox) only when the site has DataDome or Cloudflare Turnstile in front. Every production change is verified in a separate test account before promotion, and expensive fallbacks are charged only when they produce usable data. That keeps entry prices attractive without hiding the real cost of difficult sources.

The engineering writeups live on The Watchtower: anti-bot bypasses, cost teardowns, why a $0.20/result actor became a $0.008/result actor without losing reliability, and similar field notes.

Who uses Thirdwatch

• Builders & integrators — embed structured external data inside SaaS products, agents, and internal tools without owning a scraping pipeline.
• Recruiters & ATS vendors — pipe LinkedIn, Naukri, AmbitionBox, CutShort, and 15+ other job/candidate sources into hiring workflows.
• E-commerce & retail teams — track Amazon, Flipkart, Myntra, AJIO, Nykaa, Meesho, Tata Cliq, FirstCry, and AliExpress prices and stock signals.
• Researchers & VCs — hiring velocity, startup signal, supplier/distributor maps, regional food-delivery data, and trade flows.
• AI / agent teams — the MCP server is the easiest way to give an LLM agent live access to the public web.