Company Research MCP for AI Agents

An MCP server that gives Claude, ChatGPT, Gemini, Cursor and any other MCP client ten company research tools:

Tool	What the agent gets
get_website_contacts	Emails (with a check that the domain accepts mail), phone numbers, social profiles (LinkedIn, X/Twitter, Facebook, Instagram, YouTube, TikTok, GitHub...), named people, address and careers page, read on the company's own website.
identify_company	The legal company behind a European website, from its legal notice: registered name, legal form, registration number checked by its control digit (German Handelsregister, French SIREN, Italian Partita IVA, Spanish CIF, UK company number, Dutch KvK, Belgian KBO/BCE, Danish CVR, Swedish and Norwegian organisation numbers, Finnish Y-tunnus), VAT ID, register, address and managing directors. 13 countries.
find_new_us_businesses	Businesses newly registered in the US, from official registries: new LLCs, corporations and DBAs in New York, Colorado, Connecticut, Oregon and Pennsylvania, and new businesses in Los Angeles, San Francisco and Chicago, with owners, registered agent, and email and NAICS industry where published.
search_jobs	Open jobs straight from the career sites of 20,000+ companies (Greenhouse, Lever, Ashby, Workable, Recruitee, Personio, Teamtailor, Gem), indexed every day: title, company, location, remote, seniority, salary when published, posting date and the apply link.
get_tech_stack	The technologies a website runs (CMS, e-commerce platform, analytics, ads, chat, payments, frameworks, CDN) plus its email provider and the SaaS tools verified in its DNS.
search_insider_trades	Stock trades that directors, officers and 10% owners reported to the SEC on Form 4: who, buy or sell, shares, price, value, holdings after, planned or not.
search_startup_funding	Companies that just reported raising money on SEC Form D: amount raised and round size, industry, year of incorporation, address, phone and executives.
lookup_legal_entity	Any company in the global LEI system (GLEIF): LEI, exact legal name, legal form, company register and number there, addresses, status, and its direct and ultimate parent companies.
get_company_financials	Annual figures of a US-listed company from its 10-K reports at the SEC: revenue, gross profit, operating income, net income, diluted EPS, operating cash flow, assets, liabilities, equity and cash, year by year.
list_sec_filings	The latest SEC filings of a company (10-K, 10-Q, 8-K with the items it discloses, proxy statements...), with links to each document.

Ask your agent things like "Who is the company behind zalando.de, and what is its VAT ID?", "Find the sales email and LinkedIn of acme.com", "List restaurants that registered in Connecticut this month with their email" "Find remote senior data engineer jobs posted this week that publish a salary", "Is allbirds.com on Shopify, and who hosts its email?", "Did Alphabet insiders sell stock this month?" or "Which health-care startups in California raised money this week?": it picks the tool and gets clean JSON back. You pay only for results: a call that finds nothing is free.

Connect your agent

Easiest, with your Apify account (sign-in in the browser, no token to copy): add this remote MCP server URL to your client:

https://mcp.apify.com?tools=digital_influx/company-research-mcp

• Claude (claude.ai, Claude Desktop): add a custom connector with that URL.
• ChatGPT: add it as a connector (MCP server URL) in developer mode.
• Cursor, VS Code, Windsurf, Gemini CLI and other clients: add a remote MCP server. With a token instead of the browser sign-in:

{
  "mcpServers": {
    "company-research": {
      "url": "https://mcp.apify.com?tools=digital_influx/company-research-mcp",
      "headers": { "Authorization": "Bearer YOUR_APIFY_TOKEN" }
    }
  }
}

(Gemini CLI writes httpUrl instead of url in ~/.gemini/settings.json.)

Directly, without the Apify MCP server: this Actor runs in Standby mode as a Streamable HTTP MCP server. Take its Standby URL from the Actor's API tab in Apify Console, add /mcp, and send your Apify token as Authorization: Bearer YOUR_APIFY_TOKEN.

A record of every call: each tool call your agent makes is also saved as a row of the Standby run's dataset (tool, arguments, a one-line summary and the result), so you can check later what it asked and got. Without an agent: a standard run calls one tool from the input (tool and arguments) and saves the result to the dataset, so the tools also work from the API, schedules and integrations (Make, n8n, Zapier).

The tools

get_website_contacts

{ "url": "37signals.com", "maxPages": 3 }

Reads the homepage and up to maxPages contact, about, team and imprint pages (5 by default, at most 10), respecting robots.txt. Returns emails, emailDetails (where each email was found, whether it is the site's own domain, whether the domain accepts mail), phones, socials, people, address, careersPage, jobBoards, language and pagesVisited.

identify_company

{ "url": "hetzner.com" }

Real result (2026-10-04 page): Hetzner Online GmbH, GmbH, HRB 6089 (register Ansbach), VAT ID DE812871812, Industriestr. 25, 91710 Gunzenhausen, DE, managing directors Martin Hetzner, Stephan Konvickova and Günther Müller, and the legal notice page. Some big retailers refuse requests from cloud servers, where Apify runs: the tool then reports the site as blocked instead of working around it. The country comes from the domain; for .com and other generic domains the tool reads the legal notice as French, Italian, Spanish or British first, then as a German Impressum, then for Benelux and Nordic numbers, reading each page only once. Add "country": "nl" (or de, at, ch, fr, it, es, uk, be, dk, se, no, fi) to choose. Numbers with a control digit are only returned when it checks out.

find_new_us_businesses

{ "sources": ["connecticut"], "keywords": ["restaurant"], "onlyWithEmail": true, "registeredWithinDays": 30, "limit": 20 }

All arguments are optional: sources (new-york, colorado, connecticut, oregon, pennsylvania, los-angeles, san-francisco, chicago), registeredWithinDays (7 by default), entityKinds (llc, corporation, nonprofit, partnership, cooperative, assumed-name, other), keywords, cities, zipCodes, naicsPrefixes, onlyWithEmail and limit (20 by default, at most 100). Newest first, the registries taking turns within a day. Oregon and Pennsylvania publish monthly: the result says when a registry has nothing that recent.

search_jobs

{ "keywords": ["data engineer"], "remoteOnly": true, "postedWithinDays": 7, "onlyWithSalary": true, "limit": 10 }

All arguments are optional: keywords (words in the title), locations (city, country or country code), remoteOnly, seniority (intern, entry, senior, lead, director, executive, unspecified), postedWithinDays, employmentTypes, descriptionKeywords (skills the description must name, e.g. Python; slower, since descriptions are read), onlyWithSalary, onlyVisaSponsorship, companies, limit (10 by default, at most 50) and maxPerCompany (3 by default, so one big employer does not fill the answer). Newest first, without the long descriptions, so the answer stays short for the model. The tool runs the Job Search Actor of the same author in your account and takes 20 to 60 seconds.

get_tech_stack, search_insider_trades, search_startup_funding

{ "url": "allbirds.com" }

{ "tickers": ["GOOGL", "NVDA"], "filedWithinDays": 14, "limit": 20 }

{ "industries": ["health-care"], "locations": ["CA"], "filedWithinDays": 7, "limit": 20 }

search_insider_trades also takes transactionCodes (P and S, open-market purchases and sales, by default), roles, minValue and excludePlannedTrades; search_startup_funding takes industries, locations, minAmountSold and onlyYoungCompanies. Like search_jobs, these three run another Actor of the same author in your account (Tech Stack Detector, SEC Insider Trading, SEC Form D) and return a compact answer.

lookup_legal_entity, get_company_financials, list_sec_filings

{ "name": "Google LLC" }

Real result (2026-10-04): GOOGLE LLC, LEI 7ZW8QJWVPR4P1J1KQY45, Limited Liability Company, number 3582691 at the Delaware Division of Corporations, ultimate parent ALPHABET INC. Also by lei, and with country to search one country. LEIs with wrong check digits are refused before any request.

{ "company": "AAPL", "years": 4 }

Real result (10-K of 2025-10-31): fiscal year ended 2025-09-27, revenue USD 416,161,000,000, net income USD 112,010,000,000, diluted EPS 7.46, total assets USD 359,241,000,000. company takes a ticker, a CIK or the exact name as the SEC lists it. Foreign companies that report IFRS on Form 20-F have no US GAAP figures there, and cost nothing.

{ "company": "TSLA", "forms": ["8-K"], "filedWithinDays": 90 }

Each 8-K comes with the items it discloses, named as the SEC names them (2.02 Results of operations, 5.02 Departure or appointment of officers...). Insider forms 3, 4, 5 and 144 are left out unless you name them in forms.

How much it costs

Pay per event, only for results:

Event	When	Price
contacts	get_website_contacts found contact details	USD 0.02
company	each company identify_company or lookup_legal_entity identifies	USD 0.01
business	each business find_new_us_businesses returns	USD 0.004
financials	get_company_financials returned a company's annual figures	USD 0.02
filings	list_sec_filings listed filings	USD 0.01

search_jobs, get_tech_stack, search_insider_trades and search_startup_funding charge nothing here: the Actor each one runs bills its own events (Job Search: USD 0.01 per search plus USD 0.006 per job; Tech Stack Detector: USD 0.003 per website; SEC Insider Trading: USD 0.002 per transaction; SEC Form D: USD 0.005 per filing). A website without contact details or without a company number, a blocked site, or a search with no results costs nothing. Runs in Standby use 256 MB.

Good to know

• Respectful by design: an honest User-Agent (company-research-mcp), robots.txt read for this Actor's own name on every site, one page at a time per site, no proxies, and a site that refuses robots (403) is reported, never worked around.
• Results are what the sources publish. Legal notices are mandatory in the EU and the UK, so most company sites have one; sites that build it with JavaScript only, or put it in a PDF, come back as no_data.
• US registries: official open data of the Secretaries of State and the cities, read through their Socrata APIs one request a second. Emails and phones are the ones the registries publish: follow CAN-SPAM, the GDPR where it applies, and the TCPA and state rules for calls and texts.
• The same engines as the single-purpose Actors of this author: Website Contact Scraper, German Impressum Scraper, Legal Notice Extractor, Company ID Finder (Benelux and Nordics) and New Business Leads, if you want to run them on thousands of sites in one batch.
• Job search for agents: Job Search MCP by the same author lists every open job of a company live and reads full job descriptions, on top of the job search.

Sources and attribution

The website tools read the public pages of each site. get_company_financials and list_sec_filings read the SEC's EDGAR data (data.sec.gov), at the pace the SEC asks for, with a User-Agent that names this Actor. lookup_legal_entity reads GLEIF's public API; LEI data is published by GLEIF under CC0. find_new_us_businesses uses: New York State Department of State (data.ny.gov, OPEN-NY Terms of Use); Colorado Department of State (data.colorado.gov, public domain); Connecticut Secretary of the State (data.ct.gov, public domain); Oregon Secretary of State Corporation Division (data.oregon.gov, public domain); Pennsylvania Department of State (data.pa.gov, public domain); City of Los Angeles Office of Finance (data.lacity.org, CC0); San Francisco Treasurer & Tax Collector (data.sf.gov, PDDL); and the City of Chicago Department of Business Affairs and Consumer Protection (Chicago Data Portal). The Chicago data has been modified for use from its original source, www.cityofchicago.org; the City of Chicago makes no claims as to its content, accuracy, timeliness or completeness. No state or city endorses this Actor or the uses of its results.