\n\nConnect LLMs and AI agents to live web data using MCP (Model Context Protocol). The Decodo MCP\nServer lets you scrape websites, search engines, eCommerce platforms, and social media directly from\nAI tools like Claude, Cursor, and Windsurf, all without the need to build scraping infrastructure\nfrom scratch.\n\n- Structured outputs in JSON, Markdown, and screenshots\n- Server-side JavaScript rendering and anti-bot handling\n- 125M+ IPs across 195+ locations\n\n## What is Decodo MCP server?\n\nThe Decodo MCP Server is a web scraping layer for AI agents. It connects MCP-compatible clients to\nDecodo's Web Scraping API, enabling:\n\n- Web scraping for LLMs\n- Real-time data retrieval for RAG\n- AI agent browsing and research\n- Structured data extraction from dynamic websites\n\nInstead of maintaining proxies, parsers, and retry logic, you get a single integration point for\nreliable web data access.\n\n## Why use MCP for web scraping?\n\nModel Context Protocol (MCP) is the emerging standard for connecting AI agents to external tools and\ndata sources. With MCP:\n\n- Agents can call tools dynamically\n- Integrations stay standardized\n- Workflows scale across environments\n\nThe Decodo MCP Server gives your agents reliable, production-ready web access through this standard.\n\n## Key features\n\n**Web scraping for AI agents, no infrastructure required.** Scrape any website, including\nJavaScript-heavy pages, without handling proxy rotation, CAPTCHA solving, or anti-bot systems.\n\n**Structured outputs for LLM workflows.** Markdown (LLM-ready), JSON (for structured pipelines), and\nscreenshots (for visual context), built for RAG pipelines, AI research agents, and automation flows.\n\n**Built-in support for popular targets.** Ready-made tools for Google and Bing (SERPs), Amazon,\nWalmart, and Target (eCommerce), TikTok and YouTube (social media), and ChatGPT and\nPerplexity (AI search).\n\n**Global proxy infrastructure.** 115M+ residential IPs, 195+ geo-locations, and a 99.99% success\nrate on even the most protected targets.\n\n**Modular MCP toolsets.** Enable only what you need: `web`, `search`, `ecommerce`, `social_media`,\n`ai` for cleaner tool selection and better agent performance.\n\n**Fast time to value.** From API key to first scrape in minutes, no setup overhead.\n\n## Use cases\n\nUse the Decodo MCP Server when you need web scraping for AI agents, structured data extraction at\nscale, reliable access to dynamic websites, real-time data for RAG, or an alternative to building\nscraping infrastructure from scratch. Common scenarios:\n\n- **AI-powered web scraping** – give LLMs the ability to collect fresh data instead of relying on\n static training data.\n- **RAG with live data** – pull real-time Google, Bing, and AI search results into retrieval\n pipelines.\n- **eCommerce intelligence** – track product prices, listings, and sellers across marketplaces\n without getting blocked.\n- **Social media data collection** – gather posts, channels, and engagement data from TikTok\n and YouTube.\n- **Travel and price aggregation** – build tools that collect live pricing and availability across\n websites.\n\n## Quick start\n\n1. **Create a free account** at [dashboard.decodo.com](https://dashboard.decodo.com/) – up to 10K\n free requests, no credit card required.\n2. **Get your API key.** Copy your Web Data API key from your Web Data API subscription on the\n [dashboard](https://dashboard.decodo.com/web-data/playground). Older plans have a basic\n authentication token, which also works (see [Authentication](#authentication)).\n3. **Download Node.js 18+** from https://nodejs.org.\n4. **Get MCP client** like Claude Desktop, Cursor, Windsurf or other MCP-compatible tools.\n5. **Configure the MCP server** in your AI client (see configuration examples below).\n\n## Connecting to [Decodo's MCP server](https://mcp.decodo.com/mcp)\n\nOpen your preferred MCP client and add the following configuration (see examples for Claude Code,\nCursor, Windsurf bellow):\n\n```\n{\n \"mcpServers\": {\n \"Decodo\": {\n \"url\": \"https://mcp.decodo.com/mcp\",\n \"headers\": {\n \"Authorization\": \"Bearer \"\n }\n }\n }\n}\n```\n\n### Authentication\n\n| Credential | Where to get it | Hosted server header | Local server env |\n| -------------------------------------------- | ---------------------------------------------------------------------------------------------- | ----------------------------------------- | -------------------------------------- |\n| **API key** (recommended) | Web Data API subscription on the [dashboard](https://dashboard.decodo.com/web-data/playground) | `Authorization: Bearer ` | `SCRAPER_API_TOKEN=` |\n| **Basic authentication token** (older plans) | Web Scraping API page on the dashboard | `Authorization: Basic ` | `SCRAPER_API_TOKEN=` |\n\nThe local server reads either credential from `SCRAPER_API_TOKEN` and works out which one it got.\n\n### Claude Desktop\n\n1. Open Claude Desktop → Settings → Developer → Edit Config.\n2. Add to claude_desktop_config.json:\n\n```\n{\n \"mcpServers\": {\n \"Decodo\": {\n \"command\": \"npx\",\n \"args\": [\"-y\", \"@decodo/mcp-server\"],\n \"env\": {\n \"SCRAPER_API_TOKEN\": \"\",\n \"TOOLSETS\": \"web,ai\"\n }\n }\n }\n}\n```\n\n3. Save and restart Claude Desktop.\n\n### Cursor\n\n1. Open Settings → MCP.\n2. Click Add a new global MCP server (opens mcp.json).\n3. Add the same configuration as above.\n4. Save — look for a green status indicator next to Decodo.\n\n### Windsurf\n\n1. Open Settings → Windsurf Settings.\n2. Scroll to Cascade → Add custom server + (opens mcp_config.json).\n3. Add the same configuration as above.\n4. Save and restart Windsurf.\n\n## Test your setup\n\nOnce connected, try this prompt in your client:\n\n▎ \"Scrape the titles of the top 5 articles from Hacker News\"\n\nYou should get a structured list back within seconds. If you see an auth error, double-check your\nAPI key on the dashboard.\n\n## Local development\n\nOnly needed if you want to change the server code.\n\n\nBuild and run the server from this repository\n\n### Prerequisites\n\n- Node.js 18.0+\n- An MCP client - popular choices are [Claude Desktop](https://claude.ai/download) and\n [Cursor](https://www.cursor.com/)\n\n### Step-by-step guide\n\n1. Clone this repository:\n\n```\ngit clone https://github.com/Decodo/mcp-server\n```\n\n2. Run the following commands in the terminal:\n\n```\ncd mcp-server\nnpm install\nnpm run build\n```\n\n3. Take note of your build location:\n\n```\ncd build/\npwd\n```\n\nAdding `index.js` to the end of this directory, your build file location should look something like\nthis:\n\n```\n/Users/your.user/projects/mcp-server/build/index.js\n```\n\n4. Update your MCP client with the server information:\n\n```\n{\n \"mcpServers\": {\n \"decodo-mcp\": {\n \"command\": \"node\",\n \"args\": [\"/Users/your.user/projects/mcp-server/build/index.js\"],\n \"env\": {\n \"SCRAPER_API_TOKEN\": \"\"\n }\n }\n }\n}\n```\n\n\n\n## Toolsets\n\nTools are organized into toolsets. Which ones you get depends on how you connect.\n\n### Claude (official connector)\n\nThe Decodo connector in Claude's connector directory includes the Web Data API tools only:\n\n| Toolset | Tools |\n| -------- | --------------------------------------------------------------------------------------------------------- |\n| `web` | `scrape_as_markdown`, `screenshot`, `map`, `crawl_start`, `crawl_status`, `crawl_results`, `crawl_cancel` |\n| `search` | `search` |\n\n### Any MCP client\n\nYou can selectively enable specific toolsets by passing a comma-separated list via the `toolsets`\nquery parameter:\n\n```\n \"Decodo MCP Server\": {\n \"url\": \"https://mcp.decodo.com/mcp?toolsets=web,ai\",\n \"headers\": {\n \"Authorization\": \"Bearer \"\n }\n }\n```\n\nWhen no toolsets are specified, all tools are registered.\n\n| Toolset | Tools |\n| -------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |\n| `web` | `scrape_as_markdown`, `screenshot`, `map`, `crawl_start`, `crawl_status`, `crawl_results`, `crawl_cancel` |\n| `search` | `search`, `google_search`, `google_ads`, `google_lens`, `google_travel_hotels`, `bing_search` |\n| `ecommerce` | `amazon_search`, `amazon_product`, `amazon_pricing`, `amazon_sellers`, `amazon_bestsellers`, `walmart_search`, `walmart_product`, `target_search`, `target_product`, `tiktok_shop_search`, `tiktok_shop_product`, `tiktok_shop_url` |\n| `social_media` | `tiktok_post`, `youtube_metadata`, `youtube_channel`, `youtube_subtitles`, `youtube_search` |\n| `ai` | `chatgpt`, `perplexity`, `google_ai_mode` |\n\n## Tools\n\nThe server exposes the following tools:\n\n| Tool | Description | Example prompt |\n| ---------------------- | ------------------------------------------------------------------------------------------ | --------------------------------------------------------------------------------------- |\n| `scrape_as_markdown` | Scrapes any target URL, expects a URL to be given via prompt. Returns results in Markdown. | Scrape peacock.com from a US IP address and tell me the pricing. |\n| `screenshot` | Captures a screenshot of any webpage and returns it as a PNG image. | Take a screenshot of github.com from a US IP address. |\n| `map` | Lists the URLs on a website without fetching their content. API key only. | Map docs.decodo.com and list the pages under /scraping-api. |\n| `crawl_start` | Starts crawling a website and returns the crawl id without waiting. API key only. | Crawl the first 10 pages of decodo.com/blog. |\n| `crawl_status` | Reports a crawl's progress: pages discovered, completed and failed. API key only. | Is my crawl finished yet? |\n| `crawl_results` | Returns crawled pages as Markdown, a few at a time with a cursor. API key only. | Show me the pages my crawl found. |\n| `crawl_cancel` | Stops a running crawl; pages crawled so far stay readable. API key only. | Cancel my crawl. |\n| `search` | Searches the web and returns ranked results. API key only. | Search the web for MCP server hosting options. |\n| `google_search` | Scrapes Google Search for a given query, and returns parsed results. | Scrape Google Search for shoes and tell me the top position. |\n| `google_ads` | Scrapes Google Ads search results. | Scrape Google Ads for laptop and show me the top ads. |\n| `google_lens` | Scrapes Google Lens image search results. | Search Google Lens for this image: https://example.com/image.jpg |\n| `google_ai_mode` | Scrapes Google AI Mode (Search with AI) results. | Ask Google AI Mode: What are the top three dog breeds? |\n| `google_travel_hotels` | Scrapes Google Travel Hotels search results. | Search Google Travel Hotels for hotels in Paris. |\n| `amazon_search` | Scrapes Amazon Search for a given query, and returns parsed results. | Scrape Amazon Search for wireless keyboard. |\n| `amazon_product` | Scrapes Amazon Product page. | Scrape Amazon product B09H74FXNW and show me the details. |\n| `amazon_pricing` | Scrapes Amazon Product pricing information. | Get pricing for Amazon product B09H74FXNW. |\n| `amazon_sellers` | Scrapes Amazon Seller information. | Get information about Amazon seller A1R0Z7FJGTKESH. |\n| `amazon_bestsellers` | Scrapes Amazon Bestsellers list. | Show me Amazon bestsellers in electronics. |\n| `walmart_search` | Scrapes Walmart Search for a given query, and returns parsed results. | Scrape Walmart Search for camping tent. |\n| `walmart_product` | Scrapes Walmart Product page. | Scrape Walmart product 15296401808. |\n| `target_search` | Scrapes Target Search for a given query, and returns parsed results. | Scrape Target Search for kitchen appliances. |\n| `target_product` | Scrapes Target Product page. | Scrape Target product 92186007. |\n| `tiktok_post` | Scrapes a TikTok post URL for structured data (e.g. engagement, caption, hashtags). | Scrape this TikTok post: https://www.tiktok.com/@nba/video/7393013274725403950 |\n| `tiktok_shop_search` | Scrapes TikTok Shop Search for a given query, and returns parsed results. | Scrape TikTok Shop Search for phone cases. |\n| `tiktok_shop_product` | Scrapes TikTok Shop Product page. | Scrape TikTok Shop product 1731541214379741272. |\n| `tiktok_shop_url` | Scrapes TikTok Shop page by URL. | Scrape this TikTok Shop URL: https://www.tiktok.com/shop/s?q=HEADPHONES |\n| `youtube_metadata` | Scrapes YouTube video metadata. | Get metadata for YouTube video dFu9aKJoqGg. |\n| `youtube_channel` | Scrapes YouTube channel videos. | Scrape YouTube channel @decodo_official. |\n| `youtube_subtitles` | Scrapes YouTube video subtitles. | Get subtitles for YouTube video L8zSWbQN-v8. |\n| `youtube_search` | Search YouTube videos. | Search YouTube for \"How to care for chinchillas\". |\n| `bing_search` | Scrapes Bing Search results. | Search Bing for laptop reviews. |\n| `chatgpt` | Sends a single prompt to ChatGPT and returns its answer (no conversation history). | Ask ChatGPT to explain quantum computing in simple terms. |\n| `perplexity` | Sends a single prompt to Perplexity and returns its answer (no conversation history). | Ask Perplexity what the latest trends in web development are. |\n\n## Parameters\n\nThe following parameters are inferred from user prompts:\n\n| Parameter | Description |\n| --------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |\n| `jsRender` | Renders target URL in a headless browser. |\n| `geo` | Sets the country from which the request will originate. |\n| `locale` | Sets the locale of the request. |\n| `tokenLimit` | Maximum number of characters to return (default 45000). Useful if the context window is small. |\n| `prompt` | Prompt to send to AI tools (`chatgpt`, `perplexity`). |\n| `search` | Activates ChatGPT's web search functionality (`chatgpt` only). |\n| `xhr` | When true, includes XHR or fetch responses in the scrape result where supported (e.g. `tiktok_post`). |\n| `deviceType` | Device type to emulate (`desktop`, `mobile`, `tablet`). Parsed targets are stricter: Amazon tools are desktop-only so they omit it, and `bing_search`/`google_ai_mode` accept `desktop` and `mobile` only. |\n| `domain` | Domain to use for the request (e.g., `amazon.com`, `amazon.co.uk`, `bing.com`). |\n| `pageFrom` | Starting page number for pagination. |\n| `deliveryZip` | ZIP code for delivery location (Target, Walmart). |\n| `storeId` | Store ID for local inventory (Walmart). |\n| `country` | Country for TikTok Shop requests. |\n| `limit` | Maximum number of results to return (e.g., YouTube channel videos). |\n| `language_code` | Language code for subtitles (e.g., `en`, `es`). |\n\n## Examples\n\n### Scraping geo-restricted content\n\nQuery your AI agent with the following prompt:\n\n```\nScrape peacock.com from a German IP address and tell me the pricing.\n```\n\nThis prompt will say that peacock.com is geo-restricted. To bypass the geo-restriction:\n\n```\nScrape peacock.com from a US IP address and tell me the pricing.\n```\n\n### Limiting number of response tokens\n\nIf your agent has a small context window, the content returned from scraping will be automatically\ntruncated, in order to avoid context-overflow. You can increase the number of tokens returned within\nyour prompt:\n\n```\nScrape hacker news, return 50k tokens.\n```\n\nIf your agent has a big context window, tell it to return `full content`:\n\n```\nScrape hacker news, return full content.\n```\n\n## Use with Decodo agent skills\n\nThis server gives your agent the scraping _tools_.\n[`Decodo/agent-skills`](https://github.com/Decodo/agent-skills) teaches it **when** to reach for\nthem, **which** surface to use, and **how** to call it — so the agent picks the right tool and\nparameters on its own instead of guessing.\n\nThe skills route across the `decodo` CLI, this hosted MCP server, and the raw HTTP API, and fall\nback to the MCP server automatically when no shell is available. Pairing the two gives your agent\nboth the execution layer (this server) and the decision layer (the skills) in a single setup.\n\n**Quick start** (Claude Code):\n\n```text\n/plugin marketplace add Decodo/agent-skills\n/plugin install decodo@decodo-skills\n```\n\nSee the [agent-skills README](https://github.com/Decodo/agent-skills#readme) for manual install and\nother agents (Cursor, Codex, Gemini CLI, Windsurf).\n\n## Related repositories\n\n[Web Scraping API](https://github.com/Decodo/Web-Scraping-API),\n[Decodo agent skills](https://github.com/Decodo/agent-skills),\n[Decodo OpenClaw skill](https://github.com/Decodo/decodo-openclaw-skill)\n\n## Try it\n\nPlug Decodo MCP Server into your AI workflow in just a few clicks and equip your AI agents with\nreal-time data from any website.\n\n[Start for free](https://dashboard.decodo.com/) | [Docs](https://help.decodo.com/docs/introduction)\n| [Discord](https://discord.gg/Ja8dqKgvbZ)\n\n## License\n\nAll code is released under the [MIT License](https://github.com/Decodo/Decodo/blob/master/LICENSE).\n","readmeFilename":"README.md"}
![\"Decodo:](\"https://raw.githubusercontent.com/Decodo/mcp-server/main/assets/banner.png\")
\n