MCP Generator: Turn Any REST API into AI-Ready MCP Tools
MCP Generator is Akantik's own SaaS platform that turns existing REST APIs into Model Context Protocol (MCP) servers, so AI assistants and agents such as Claude, Cursor and GitHub Copilot can call a business's real systems as tools, without anyone writing MCP server code.
Teams import a Postman collection or an OpenAPI/Swagger spec, organise the endpoints into groups, and publish them as a hosted, OAuth2-secured MCP server with its own URL. Every AI tool call is logged, so teams always know what their agents are doing.
Project Gallery
Businesses increasingly want AI assistants that can act on their own data, such as looking up an order, updating a customer record or checking stock levels. But connecting those assistants to existing systems takes a lot of custom work. Every API had to be wrapped in a hand-written MCP server that someone then had to host and maintain, even though most teams had already documented those same APIs in Postman or OpenAPI and couldn't reuse any of it.
Security and visibility were equally difficult. Letting an AI agent call production APIs without proper authentication is a serious risk, and once an agent was connected, teams had no clear way to see which tools it called, when, or whether those calls succeeded. AI tools were also easy to misconfigure, because a vague tool name or a DELETE endpoint without a record ID could lead an agent to call the wrong thing or change the wrong record.
We designed and built MCP Generator as a self-service platform that takes a team from an existing API to a live, secured MCP server in minutes. Instead of rewriting anything, teams import a Postman Collection v2.1 or an OpenAPI 3.x / Swagger 2.0 spec, and every request becomes an endpoint inside a new group. Path parameters such as {id} automatically become tool inputs, and any items that can't be parsed are reported separately so they never break the whole import.
Each endpoint becomes an AI tool with a clear snake_case name and a plain-English description that tells the AI when to call it. Parameters, headers, authorization and request body are configured per endpoint, and a built-in Test button checks the real response before anything goes live. The editor also warns about risky setups, such as a DELETE endpoint with no record ID, before an agent can misuse it.
Groups are then published as hosted MCP servers, each with its own URL and a Draft/Live switch, ready to connect to Claude, Cursor or VS Code. OAuth2 can be required on any server so AI clients must sign in first, every tool call is logged with its status and response time, and tiered subscription plans handle usage limits, renewals and invoices.
Technical Stack
- Frontend
- Blazor HTML5 CSS3 JavaScript
- Backend
- ASP.NET Core
- Framework
- .NET Model Context Protocol (MCP)
- Architecture
- Multi-tenant SaaS
- Database
- PostgreSQL
- Cloud
- IIS
- Versioning
- Git
- Project Management
- Azure DevOps
Key Features
MCP Generator lets teams import Postman collections and OpenAPI/Swagger specs and turn them into AI tools in a single upload. A visual endpoint editor covers the method, URL, parameters, headers, authorization and body, with live testing built in. Endpoints are organised into API groups, and any group can be published as a hosted MCP server with its own URL and a Draft or Live status.
Security and control are built into every server. OAuth2 protection makes AI clients sign in before calling any tool, users can review and revoke connected AI apps at any time, and step-by-step guides cover connecting Claude Desktop, Cursor and VS Code (GitHub Copilot).
For visibility, every tool call is recorded in the MCP call logs with its status and response time, and the analytics dashboard shows daily usage, success rate, the most-used tools and calls by HTTP method. Teams can export their servers, groups and endpoints as JSON, with credentials masked, and their call logs as CSV. Subscription plans with usage-based limits, renewals and invoices are included.
Let's build something together
MCP Generator shows how Akantik turns emerging AI standards into working products. By turning existing APIs into secure, fully logged MCP tools, it lets businesses connect AI assistants to their real systems in minutes instead of weeks. If you want AI to work with your own data, securely and with full visibility, we can build that connection for you.