# Swagger to MCP Gateway

> Swagger to MCP Gateway is a secure MCP gateway for publishing OpenAPI APIs and database scopes as controlled Model Context Protocol tools for AI agents.

Swagger to MCP Gateway helps teams expose existing REST APIs and approved PostgreSQL or SQL Server database scopes to MCP-compatible clients such as Claude, Cursor, VS Code, and other AI agent runtimes. The product keeps credentials, policy checks, validation, rate limits, and audit logs inside the Gateway instead of placing them in prompts or local client configuration.

## Core Pages

- [MCP Gateway](https://swaggertomcp.com/mcp-gateway): Product overview for the secure AI-to-system gateway.
- [OpenAPI to MCP](https://swaggertomcp.com/openapi-to-mcp): Convert Swagger/OpenAPI operations into MCP tools.
- [Database to MCP](https://swaggertomcp.com/database-to-mcp): Publish PostgreSQL and SQL Server scopes as metadata-only schema tools and validated executor tools.
- [Docs](https://swaggertomcp.com/docs): Setup guide for OpenAPI sources, database scopes, MCP clients, and Gateway security.
- [Resources](https://swaggertomcp.com/resources): Technical guides for OpenAPI to MCP, Database to MCP, MCP client setup, and MCP tool security.
- [Pricing](https://swaggertomcp.com/pricing): Plans for individual and team usage.
- [About](https://swaggertomcp.com/about): Company, engineering background, and trust context.
- [Contact](https://swaggertomcp.com/contact): Sales, support, and partnership contact.

## Product Capabilities

- Convert Swagger/OpenAPI definitions into MCP tools with schema-aware arguments and editable descriptions.
- Publish PostgreSQL and SQL Server database scopes as MCP schema tools without returning database row payloads.
- Validate QueryPlan and SQL executor calls in the Gateway before database execution.
- Keep upstream API credentials and database connection strings server-side.
- Apply account limits, quotas, redaction, and usage observability to MCP tool traffic.

## Technical Guides

- [OpenAPI to MCP guide](https://swaggertomcp.com/resources/openapi-to-mcp-guide): How to turn REST APIs into agent-ready MCP tools.
- [MCP Gateway vs custom MCP server](https://swaggertomcp.com/resources/mcp-gateway-vs-custom-mcp-server): Comparison for production MCP architecture.
- [Database to MCP security](https://swaggertomcp.com/resources/database-to-mcp-security): Schema tools, scopes, and validated database execution.
- [QueryPlan vs SQL mode](https://swaggertomcp.com/resources/queryplan-vs-sql-mode): Database executor mode comparison.
- [Claude, Cursor, and VS Code MCP setup](https://swaggertomcp.com/resources/claude-cursor-mcp-setup): Connecting MCP clients to Gateway endpoints.
- [MCP tool security checklist](https://swaggertomcp.com/resources/mcp-tool-security-checklist): Checklist for production AI tool governance.
- [What is MCP?](https://swaggertomcp.com/resources/what-is-mcp): Plain-English Model Context Protocol explanation.
- [What is an MCP server?](https://swaggertomcp.com/resources/what-is-an-mcp-server): MCP server responsibilities and gateway model.
- [What is an MCP tool?](https://swaggertomcp.com/resources/what-is-an-mcp-tool): Tool schemas, descriptions, and agent-ready tool design.
- [What is an MCP client?](https://swaggertomcp.com/resources/what-is-an-mcp-client): How AI apps discover and call MCP tools.
- [MCP vs API gateway](https://swaggertomcp.com/resources/mcp-vs-api-gateway): Difference between API gateways and MCP surfaces.
- [MCP vs function calling](https://swaggertomcp.com/resources/mcp-vs-function-calling): Comparison for AI tool workflows.
- [AI agent tool use](https://swaggertomcp.com/resources/ai-agent-tool-use): How agents use external tools.
- [LLM tool calling security](https://swaggertomcp.com/resources/llm-tool-calling-security): Validation and security boundaries for tool calls.

## Security Model

The Gateway is the security boundary. LLMs, MCP clients, and orchestration frameworks may propose tool calls, QueryPlans, or SQL candidates, but the Gateway validates source ownership, published scope, table and column permissions, execution mode, limits, and sensitive value redaction before execution.

Schema tools never return database rows. They expose approved metadata such as tables, columns, keys, relationships, permissions, and allowed operations.

## More LLM Context

- [Full LLM context](https://swaggertomcp.com/llms-full.txt)
- [AI-readable product manifest](https://swaggertomcp.com/.well-known/mcp-gateway.json)