# The Company Check AI and MCP Usage Guide

The Company Check provides public preview pages for company, brand, director, pricing, and search experiences. AI agents, MCP clients, search crawlers, and research tools may use the public pages to understand publicly visible information and navigate the site.

Base URL: https://www.thecompanycheck.com
AI manifest: https://www.thecompanycheck.com/ai-manifest.json
MCP discovery: https://www.thecompanycheck.com/.well-known/mcp.json
MCP endpoint: https://www.thecompanycheck.com/mcp

## MCP Support

The public MCP endpoint is read-only and public-preview only. It supports JSON-RPC over HTTP POST for:

- initialize
- ping
- resources/list
- resources/read
- tools/list
- tools/call

It exposes route discovery, usage guidance, and structured public profile data through these read-only tools:

- get_company_profile: public company facts by CIN or LLPIN
- get_brand_profile: public brand facts by The Company Check brand ID
- get_director_profile: public director and company-affiliation facts by DIN
- get_public_routes: canonical public route catalog
- get_page_guidance: extraction and access guidance for a canonical page URL

Profile tools return both MCP `structuredContent` and a JSON text fallback. They do not expose locked, paid, account-only, checkout, report-only, private API, residential, contact, PAN, birth, remuneration, or full shareholding data.

For company profiles, exchange-filed annual and quarterly financials are returned only when that company's listed-financial section is unlocked on its public page. If the page keeps financials locked, the MCP response contains no financial values and identifies `financial-statements` in `access.lockedSections`.

For brand profiles, MCP returns the same public summary and preview rows shown on the page: identity, location, sectors, business models, the primary legal entity, IPO listing, up to five named products, public people previews, funding summary and preview rounds, up to three public investor previews, acquisition previews, public key metrics, up to six competitor previews, and up to three news items. Complete lists, private contacts, round amounts and valuations, comparison data, and full financial statements remain locked.

For director profiles, MCP returns DIN status, nationality, the page-consistent current and former company lists, role dates, summary counts, and the public directorship timeline. Inactive, struck-off, amalgamated, dissolved, liquidated, converted, and defunct companies are classified as former roles, matching the rendered page. Shareholding, remuneration, contact, email, residential, birth, and PAN values are never returned.

## Important Routes

- Company profile: /company/{company-name}/{cin}
- Brand profile: /company/b/{brand-name}/{brand-id}
- Director profile: /people-profile/{director-name}/{din}
- Director timeline: /people-profile/{director-name}/directorship-timeline/{din}
- Director shareholding/associated-company shares: /people-profile/{director-name}/associated-company-shares/{din}
- Pricing: /pricing
- Search: /textsearch?q={query}

## Page Context

Public pages expose a JSON script block with id `tcc-ai-page-context`. It identifies the page type, entity type, entity id, route pattern, canonical URL, access state, data sources, and available sections.

Agents should read that context before extracting page content. The context is descriptive and does not grant access to locked or paid data.

## Access Rules

- Use only public, visible content unless you have an authorized account and user permission.
- Do not attempt to bypass locks, paywalls, login, checkout, Cloudflare, rate limits, or API access controls.
- Do not scrape at high rates or parallelize requests aggressively.
- Prefer canonical URLs.
- Respect robots directives and cache responsibly.
- Payment, login, report, and user-account flows must not be automated without explicit user authorization.

## Good Crawlers and AI Agents

Known beneficial crawlers and AI agents should use stable identification, reasonable request rates, and public routes. The site may challenge, throttle, or block anonymous automation, abusive IPs, and non-human traffic that harms normal users.

## Contact

For approved data access, commercial integrations, or crawler allowlisting, contact support@thecompanycheck.com.