Hexadata

Hexadata MCP — official data sources as tools for AI agents

Hexadata publishes Apify Actors that turn official registers and public portals into clean JSON. Every Actor is also an MCP tool: any MCP client (Claude Desktop, Claude Code, Cursor, VS Code, Windsurf, custom agents) can call it through the Apify MCP server, with no code to deploy.

This repository holds the connection instructions, a ready-to-use server.json for MCP registries, and example prompts. The tools themselves run on Apify and are billed per result (no results, no cost).

Tools

All tools share the same field conventions (id, title/name, url, country, publishedAt, scrapedAt), an onlyNewSinceLastRun mode for scheduled runs, and a maxResults cap that bounds the cost of every call.

Connect in one step

Add the Apify MCP server to your client with the tools you want. Use one Actor:

https://mcp.apify.com?tools=hexadatalabs/eu-tenders-ted-scraper

or several, comma-separated:

https://mcp.apify.com?tools=hexadatalabs/eu-tenders-ted-scraper,hexadatalabs/sam-gov-opportunities-scraper,hexadatalabs/uk-companies-house-scraper

Authentication: sign in with OAuth on first use (the client opens apify.com), or send your Apify token as Authorization: Bearer . Running tools needs an Apify account; a free account includes monthly credit.

Claude Desktop, Cursor, Windsurf (JSON config)

{

"mcpServers": { "hexadata": { "url": "https://mcp.apify.com?tools=hexadatalabs/eu-tenders-ted-scraper,hexadatalabs/sam-gov-opportunities-scraper,hexadatalabs/us-healthcare-providers-npi-scraper,hexadatalabs/uk-companies-house-scraper,hexadatalabs/mexico-business-directory-denue,hexadatalabs/occ-mundial-jobs-scraper,hexadatalabs/seccion-amarilla-business-scraper", "headers": { "Authorization": "Bearer " } } } }

Claude Code

claude mcp add --transport http hexadata "https://mcp.apify.com?tools=hexadatalabs/eu-tenders-ted-scraper,hexadatalabs/uk-companies-house-scraper" --header "Authorization: Bearer <APIFY_TOKEN>"

VS Code (.vscode/mcp.json)

{

"servers": { "hexadata": { "type": "http", "url": "https://mcp.apify.com?tools=hexadatalabs/sam-gov-opportunities-scraper", "headers": { "Authorization": "Bearer ${input:apify-token}" } } } }

The configurator at mcp.apify.com generates the snippet for any client; the Apify CLI can also install it (apify mcp install cursor).

Example prompts

• "Which contract notices for road maintenance were published on TED in Poland and Czechia this week, and which close before December 15?" → eu-tenders-ted-scraper
• "Find small-business set-aside solicitations for IT services in Virginia posted since Monday." → sam-gov-opportunities-scraper
• "List the physical-therapy clinics in ZIP codes starting with 787 with phone numbers." → us-healthcare-providers-npi-scraper
• "Which software companies were incorporated in Manchester in the last 7 days, and who are their directors?" → uk-companies-house-scraper (pass your Companies House key in apiKey)
• "Give me 50 dental clinics in Monterrey with coordinates." → mexico-business-directory-denue

Every Actor description and input field starts with "Use this when…", so agents pick the right tool and fill the input without extra prompting. Set maxResults in the prompt ("at most 100") to bound the cost.

From code instead of MCP

The same Actors are available through the Apify API and client libraries:

from apify_client import ApifyClient

client = ApifyClient(" ") run = client.actor("hexadatalabs/eu-tenders-ted-scraper").call(run_input={"searchQueries": ["software"], "countries": ["DE"], "maxResults": 100}) items = client.dataset(run["defaultDatasetId"]).list_items().items

Each Actor's README has curl, Python and Node examples plus scheduling and webhook notes.

Legal and data

Hexadata Actors use official APIs or public portals, respect their terms, and exclude personal data unless you opt in (for example includeOfficers on Companies House or includeContacts on SAM.gov). Attribution to the source is included in every item (source).

Support

Open an issue on the relevant Actor's Issues tab on Apify, or here. Issues are answered within 24 hours; confirmed bugs are fixed within 48 hours.