Documentation Home

Welcome to the MCP Functions documentation. Learn how to create, deploy, and manage MCP tools with our AI-powered platform.

Platform highlights

• AI code generation — Streaming and agent-based generation; optional web search; generate parameters, name, and description from a prompt.

• Tool flow diagrams — Generate flow diagrams from tool definition.

• Config & secrets — Workspace secrets and OAuth2 in tool config; resolved at execution.

• MCP — initialize, tools/list, tools/call, resources/list, resources/read, prompts/list, prompts/get, logging/setLevel.

• Auth — API key (header or query), custom headers per server.

• Tool versions — Versioning and publish; view and roll back.

• Dashboard — Stats, charts, activity feed, country invocations.

• Sandbox — VM-based execution, timeouts, HTTP limits, blocked patterns.

Sections

Section	Description
Getting Started	Introduction, quick start, account setup, and business benefits
Core Concepts	Platform architecture, organizations, workspaces, MCP servers, and tools
Creating Tools	AI-powered creation, manual creation, and code examples
Execution	Execution flow, sandbox environment, and testing tools
MCP Protocol	Protocol overview, connecting servers, and tool invocation

Quick links

• Quick Start Guide — Get up and running in minutes

• Code Examples — Real-world tool examples

• Connecting AI assistants — Connect ChatGPT, Claude, and more