Guide

What MCP tools can do

Capabilities and example use cases for InstantMCP.io servers.

An InstantMCP.io server turns your sources into tools an AI client can call: search content, fetch structured data, follow links, or run actions defined by an API. The exact tool set depends on what you connect (docs, site, OpenAPI, files).

What a tool is

Each tool has a name, a short description, and a schema for arguments. When the user asks something relevant, the client can invoke that tool and use the result in the reply. InstantMCP.io generates those tools from your source so you do not hand-write protocol code.

Common capabilities

  • Search and retrieve: find the right page, section, or snippet in your docs or knowledge base
  • Answer from your product: pricing, onboarding steps, feature limits, changelog notes
  • Navigate: return canonical URLs instead of guessed links
  • API-backed actions: list, create, or update resources when the source is an OpenAPI/spec-driven backend
  • Grounded support: agents cite your content instead of inventing product behavior

Example: product docs MCP

Connect your documentation site or Markdown repo. Typical tools might look like:

  • search_docs / search_knowledge: answer “How do I reset a webhook?”
  • get_guide: pull a named guide (connect, billing, SSO)
  • list_topics: show what the knowledge base covers

Agents in Cursor or Claude can then help teammates and customers using your wording.

Example: SaaS API MCP

Connect an OpenAPI spec. InstantMCP.io can expose operations as tools, for example:

  • list_projects, get_project
  • create_invoice, get_invoice_status
  • invite_member

The agent becomes a natural-language front end to the API, with schemas that match your spec.

Example: internal playbooks

Connect Notion pages, PDFs, or a private handbook (where supported). Useful patterns:

  • Onboarding checklists for new hires
  • Incident runbooks (“what do we do if Stripe webhooks fail?”)
  • Policy lookups with links back to the source section

Designing good sources

Cleaner input produces clearer tools:

  • Prefer structured headings and short pages over one giant dump
  • Keep OpenAPI operation summaries accurate; they often become tool descriptions
  • Re-sync after major content or API changes so tools stay in sync

Next steps

  • Getting started: create your first server
  • Connect: plug the MCP URL into Cursor or Claude
  • FAQ: sources, hosting, and security