4 posts tagged with "mcp tools"

HTML vs Markdown vs Plain Text for LLM Extraction

October 11, 2026 · 19 min read

Co-Founder @ ScrapingAnt

HTML vs Markdown vs Plain Text for LLM Extraction

Choose the smallest representation that preserves the information your extraction task needs. Plain text can be sufficient for visible values. Markdown is a useful candidate when headings, tables and link destinations matter. Keep HTML when the answer depends on DOM attributes or script data that conversion removes.

We tested the same six-field extraction contract on three saved inputs, then on 15 hosted MCP returns. HTML produced correct records in each tested case. Markdown worked for the purchase-link case but returned empty arrays for the table. Local text worked for the table; hosted raw-mode text did not. These small, fixed-model results make the practical rule sharper: preserve the fields, then validate complete records before comparing cost.

Claude Code Can't Scrape JavaScript Sites. Here's the Fix.

April 12, 2026 · 6 min read

Oleg Kulyk

Co-Founder @ ScrapingAnt

Claude Code Can't Scrape JavaScript Sites. Here's the Fix.

You ask Claude Code to fetch a website and summarize it. Instead of useful content, you get back an empty <div> and a bunch of script tags. If you've used Claude Code for web scraping for more than a week, you've hit this wall. Here's why it happens and how to fix it in 30 seconds.

Build an AI Scraper with MCP and Validate Its Extracted Records

November 24, 2025 · 9 min read

Oleg Kulyk

Co-Founder @ ScrapingAnt

Build an AI Scraper with MCP and Validate Its Extracted Records

Updated 2026-09-30

Replaced the conceptual agent pipeline with five captured MCP retrievals, 15 real Claude extractions, and an offline validator. The evidence packet preserves the raw results, including a formatting failure and quarantined records. The September 29 correction removed unsupported reliability and vendor-comparison claims; this refresh supplies an executed implementation.

An MCP tool can return page content without proving that a model extracted the correct price. A JSON Schema can accept a record that contains the wrong value. Keep retrieval, extraction and acceptance as separate steps, and retain the evidence at each boundary.

This walkthrough retrieves an owned catalog through ScrapingAnt MCP, sends the saved Markdown to a pinned Claude model, and checks its records against a schema, source excerpts and an independent fixture oracle. You can replay the complete validation offline before using either API.

What Is an MCP Server? Architecture, Transports and a Working Example

July 6, 2025 · 14 min read

Oleg Kulyk

Co-Founder @ ScrapingAnt

What Is an MCP Server? Architecture, Transports and a Working Example

Updated 2026-09-15

Rewritten against the MCP specification (version 2025-06-18) with three new diagrams. The earlier text described components that are not part of the protocol and called Server-Sent Events "in development"; the spec replaced that transport with Streamable HTTP. Every protocol statement below links to the spec, the transcripts come from real runs against ScrapingAnt's server, and every code block is a file that was executed. Code and captured output: scrapingant-examples/examples/what-is-mcp-server.

An MCP server is a program that offers tools, data and prompt templates to an AI assistant through a standard protocol, the Model Context Protocol. The assistant's application (Claude Desktop, Cursor, VS Code, Claude Code) connects to it, asks what it can do, and calls its tools when the model decides they are needed. The server can be a local process on your machine or a remote service such as ScrapingAnt's, which turns "fetch this page" into a tool the assistant can use.