Everything you need to ship MCP tools.
Quickstarts, API references, and deployment playbooks.
LegacyAI turns OpenAPI and Swagger specs into Claude Desktop-ready MCP servers. Use the docs to configure auth, safety defaults, and dynamic sync in minutes.
What is MCP
MCP (Model Context Protocol) is the open standard that lets Claude connect to your APIs and tools. LegacyAI generates MCP servers from your OpenAPI spec automatically.
Go from OpenAPI to MCP in three steps.
Connect your OpenAPI spec
Upload Swagger or OpenAPI 2.0/3.x specs, or link your live API endpoint for continuous sync.
Generate MCP tools instantly
LegacyAI builds typed MCP tools with validation, rate limits, and safe defaults in minutes.
Launch with Claude Desktop
Register the server, configure auth scopes, and ship AI-ready tools to your teams.
Core concepts to master MCP tool design.
Authentication & scopes
Map OAuth, JWT, or API key flows to MCP tool permissions with granular scope control.
Tool safety & read-only modes
Protect mission-critical systems by enforcing read-only defaults and guarded writes.
Schema validation
Automatic Zod validation keeps input and output contracts strict and reliable.
Sync & versioning
Avoid spec drift with automated sync, change alerts, and environment pinning.
API references and integration guides.
Answers for MCP setup and rollout.
How do I generate my first MCP server?
Upload your OpenAPI or Swagger spec in the LegacyAI dashboard, then select the environment and auth strategy. LegacyAI generates a production-ready MCP server with validated tools. Register it with Claude Desktop and test a read-only tool to validate responses before expanding access.
What OpenAPI versions are supported?
LegacyAI supports OpenAPI 3.x and Swagger 2.0. The platform normalizes older specs so tools are generated with consistent schemas. For best results, include request body definitions, response schemas, and explicit security schemes in the spec.
Can I use LegacyAI without exposing my API publicly?
Yes. You can upload a local spec file, export it during CI, or use the CLI to generate MCP tools. Your production API does not need to expose the spec publicly, and the MCP server runs inside your own environment.
How do I configure Claude Desktop after generating?
Download the MCP server output and add it to your Claude Desktop configuration file. Use environment variables for tokens and base URLs. Once the server is registered, refresh tools in Claude Desktop and run a read-only test call.
What happens when my API spec changes?
LegacyAI can re-sync tools automatically on server startup, or you can regenerate them in CI when API files change. This keeps the MCP tool catalog aligned with the latest schema and avoids runtime mismatches.
Need help shipping MCP tools?
Our team can help with onboarding, API reviews, and enterprise deployment planning. Reach out for tailored MCP guidance.
- Claude Desktop configuration scripts and JSON templates.
- OpenAPI validation with automatic error reporting.
- Audit logs and monitoring dashboards for MCP usage.
- Migration guides for legacy Swagger specifications.
Talk with the LegacyAI team
Share your OpenAPI spec and we'll map out the MCP rollout.