Apify Scrapers MCP server

<!-- mcp-name: io.github.jlucasmcrell/apify-scrapers --> mcp-name: io.github.jlucasmcrell/apify-scrapers

Apify Public Data MCP: 40 Production Data Tools & Public Records

One MCP server gives an AI agent 40 production data extractors - Google Maps business leads, Google News, Google Autocomplete, LinkedIn and Glassdoor jobs, SEC EDGAR, USAspending, Grants.gov, TED tenders, NHTSA recalls, FEC, EPA, ClinicalTrials.gov, openFDA, Europe PMC, GLEIF, CMS providers, Census geocoding, state business registries, contractor licences, Airbnb, YouTube, Twitch and Google Play - each running as an Actor on your own Apify account. Install with uvx apify-data-scrapers (Claude Desktop, Cursor, any MCP client), or call the same Actors directly from Python, Node.js or no-code tools.

Each actor is built with strict schema validation, deterministic field mapping, self-healing DOM selectors, and pay-per-event pricing (per-result rates from $0.00015, Actor-start from $0.0005; the live Apify Store price is authoritative).

Runtime contract and spending control

Since version 1.1.0, calls return {results, status, run?} in both MCP structured content and text. Status is success, partial, empty_unverified, or error; errors include a stable code and retry guidance. Zero rows are not proof that no matching records exist. Existing integrations that parsed a bare result array must now read results. Native dataset field names are preserved.

Each call requests a 120-second Actor timeout and a pay-per-event charge cap of $1 by default. Set APIFYMAXCHARGE_USD to adjust the cap. This is per run, not a total account spending limit; platform fees outside Actor event charges may still apply. Up to four calls may run concurrently. Cancelling a request attempts to abort its cloud run; check the returned run ID if confirmation fails. An ambiguous start is never retried automatically.

The MCP executable is Python; Node.js files are direct Actor API examples, not a second MCP server. Private queued Actors are omitted from this public catalog.

Quick Navigation

• Make.com templates
• Use-case guides
• Available Extractors & Store Listings
• Python Quickstart
• Node.js Quickstart
• No-Code & Automation Workflows (n8n, Sheets, Slack)
• Pre-Built Example Tasks (Zero Code)
• Free Sample Datasets
• AI Agent & MCP Integration (Claude Desktop, Cursor, Custom Agent)
• In-Depth Engineering Guides
• Repository Structure
• Contributing & Author

Choose your tools

Choose a focused tool profile and generate your installation configuration.

Shortened. The full README is on GitHub.