Skip to content
Back to search

MCPify

MCPify is an AI enablement compiler that transforms existing applications into AI-native, agent-operable systems.

Safety scan

A static text scan, not a guarantee. Select the badge to see what was scanned and every finding.

Checked today
Badge for your README
findsafeskills safety rating

Markdown

[![findsafeskills safety rating](/api/listings/amarnath3003--mcpify/badge.svg)](/listings/amarnath3003--mcpify)

HTML

<a href="/listings/amarnath3003--mcpify"><img src="/api/listings/amarnath3003--mcpify/badge.svg" alt="findsafeskills safety rating" /></a>

Install

i

Install instructions from the README

As written by the repo's author. findsafeskills has not verified these steps.

The easiest way to use MCPify is via `npx`. No installation required:

npx mcpify-cli analyze ./my-app

To run it against the flagship ecommerce example in this repo:

npx mcpify-cli analyze ./examples/ecommerce-saas \
  --output ./examples/ecommerce-saas/.mcpify \
  --prisma ./examples/ecommerce-saas/prisma/schema.prisma \
  --swagger ./examples/ecommerce-saas/openapi.json

Alternatively, to build from source:

git clone https://github.com/amarnath3003/MCPify.git
cd MCPify
npm install
npm run build
npm run mcpify -- analyze ./examples/ecommerce-saas \
  --output ./examples/ecommerce-saas/.mcpify \
  --prisma ./examples/ecommerce-saas/prisma/schema.prisma \
  --swagger ./examples/ecommerce-saas/openapi.json

After generation:

cd examples/ecommerce-saas/.mcpify
npm install
npm run build

The generated `AGENTS.md` explains how to connect the compiled server to an MCP client. The ecommerce walkthrough lives in `examples/ecommerce-saas/DEMO.md`.

What was scanned

The files the safety scan read, as written by the repository's author and shown as plain text. Lines with a finding are marked.

README.md · README · 242 lines

View this file on GitHub (read at the repository's current version: the scan did not record which commit it read)

  1. <p align="center">
  2. <img src="assets/logo.png" alt="MCPify" width="720" />
  3. </p>
  4. <p align="center">
  5. <strong>Compile software into AI-operable systems.</strong>
  6. </p>
  7. <p align="center">
  8. MCPify is the AI Enablement Compiler. Transform applications, APIs, frontends, workflows, and databases into AI-native systems for autonomous agents.
  9. </p>
  10. <p align="center">
  11. <a href="https://www.npmjs.com/package/mcpify-cli">
  12. <img src="https://img.shields.io/npm/v/mcpify-cli?color=FCD269" alt="npm version" />
  13. </a>
  14. <a href="https://github.com/amarnath3003/MCPify/actions">
  15. <img src="https://img.shields.io/badge/build-passing-FF8383?logo=github" alt="build status" />
  16. </a>
  17. <a href="https://www.npmjs.com/package/mcpify-cli">
  18. <img src="https://img.shields.io/npm/dm/mcpify-cli?color=9A72DE" alt="npm downloads" />
  19. </a>
  20. <a href="https://www.npmjs.com/package/mcpify-cli">
  21. <img src="https://img.shields.io/node/v/mcpify-cli?color=4FA2FB" alt="node version" />
  22. </a>
  23. <a href="https://www.npmjs.com/package/mcpify-cli">
  24. <img src="https://img.shields.io/npm/l/mcpify-cli?color=8DE344" alt="license" />
  25. </a>
  26. <a href="https://github.com/amarnath3003/MCPify/stargazers">
  27. <img src="https://img.shields.io/github/stars/amarnath3003/MCPify?color=FC9D59" alt="stars" />
  28. </a>
  29. </p>
  30. <p align="center">
  31. <a href="#overview">Overview</a> ·
  32. <a href="#features">Features</a> ·
  33. <a href="#getting-started">Getting Started</a> ·
  34. <a href="#cli">CLI</a> ·
  35. <a href="#architecture">Architecture</a>
  36. </p>
  37. ---
  38. <p align="center">
  39. <img src="https://github.com/amarnath3003/MCPify/raw/main/assets/recgif.gif"
  40. width="1080"
  41. alt="Demo recording" />
  42. </p>
  43. ---
  44. ## Overview
  45. ### Problem Statement
  46. Modern software is built for humans, not agents. The useful actions are scattered across frontend interactions, backend services, APIs, databases, and workflows, which forces AI systems into brittle browser automation or hand-written MCP boilerplate.
  47. - manual MCP tool authoring for capabilities that already exist
  48. - brittle browser automation for UI interactions
  49. - raw endpoints instead of meaningful workflows
  50. - permission and safety gaps at the tool boundary
  51. - schema and tool drift as the application changes
  52. ### Solution
  53. MCPify acts as a compiler for your application. It scans the parts of the codebase that matter to agents and produces a runnable MCP server, semantic workflows, permission-aware tools, and the metadata an agent needs to use them safely.
  54. Stop hand-writing MCP tools. Compile your stack once. Stay in sync forever.
  55. ### Codex Usage
  56. Codex is a natural fit for MCPify in a few ways:
  57. 1. **Ideate** - use Codex to explore agent workflows, sketch MCP surfaces, and refine how a product should be exposed to AI.
  58. 2. **Build** - use Codex to implement or extend the compiler, generators, analyzers, and demo apps in this repo.
  59. 3. **Run as MCP inside Codex** - connect the generated MCP server so Codex can call real tools, inspect app surfaces, and operate against the compiled environment.
  60. ## Features
  61. - **Backend Analyzer:** Deep AST analysis of routes, controllers, and services to surface every callable action. Framework-aware for Express, Fastify, NestJS, and Next.js — including inline, non-exported handlers.
  62. - **Frontend Action Extraction:** React, Vue, Svelte components mapped to agent-controllable actions.
  63. - **OpenAPI → MCP:** Drop in a spec, ship a typed MCP server in seconds.
  64. - **Workflow Engine:** Multi-step processes detected and exposed as atomic agent capabilities.
  65. - **Permission Layer:** Scopes, roles, and audit trails enforced at the tool boundary.
  66. - **AI Metadata Enhancement:** Auto-generated descriptions, hints, and examples agents actually understand.
  67. - **Database Intelligence:** Schemas, relations, and constraints become safe, queryable surfaces.
  68. - **Event System Integration:** Webhooks, queues, and pub/sub plugged into agent loops.
  69. - **Knowledge Graph Engine:** Entities, intents, and relations modeled across your stack.
  70. - **Self-Updating Sync:** MCP definitions regenerate on every commit. No drift.
  71. - **AI Simulations:** Run agents against your app in a sandbox before shipping.
  72. ## Getting Started
  73. The easiest way to use MCPify is via `npx`. No installation required:
  74. ```bash
  75. npx mcpify-cli analyze ./my-app
  76. ```
  77. To run it against the flagship ecommerce example in this repo:
  78. ```bash
  79. npx mcpify-cli analyze ./examples/ecommerce-saas \
  80. --output ./examples/ecommerce-saas/.mcpify \
  81. --prisma ./examples/ecommerce-saas/prisma/schema.prisma \
  82. --swagger ./examples/ecommerce-saas/openapi.json
  83. ```
  84. Alternatively, to build from source:
  85. ```bash
  86. git clone https://github.com/amarnath3003/MCPify.git
  87. cd MCPify
  88. npm install
  89. npm run build
  90. npm run mcpify -- analyze ./examples/ecommerce-saas \
  91. --output ./examples/ecommerce-saas/.mcpify \
  92. --prisma ./examples/ecommerce-saas/prisma/schema.prisma \
  93. --swagger ./examples/ecommerce-saas/openapi.json
  94. ```
  95. After generation:
  96. ```bash
  97. cd examples/ecommerce-saas/.mcpify
  98. npm install
  99. npm run build
  100. ```
  101. The generated `AGENTS.md` explains how to connect the compiled server to an MCP client. The ecommerce walkthrough lives in `examples/ecommerce-saas/DEMO.md`.
  102. ## CLI
  103. ### `analyze [path]`
  104. Default command. Runs the full pipeline:
  105. - backend analysis
  106. - optional OpenAPI, Prisma, Drizzle, and Mongoose analysis
  107. - event and webhook discovery
  108. - optional frontend extraction
  109. - workflow detection
  110. - permission classification
  111. - MCP server generation
  112. ```bash
  113. npx mcpify-cli analyze . \
  114. --swagger ./tests/fixtures/swagger/petstore.yaml \
  115. --prisma ./tests/fixtures/prisma/simple.prisma \
  116. --watch
  117. ```
  118. Useful flags:
  119. - `--output <dir>` change output directory (default: `./.mcpify`)
  120. - `--no-frontend` skip UI action extraction
  121. - `--no-events` skip webhook and listener analysis
  122. - `--no-workflows` skip workflow detection
  123. - `--reachable-only` emit only externally-reachable backend actions (skip internal helpers)
  124. - `--ai-enhance` improve tool descriptions (requires `ANTHROPIC_API_KEY`)
  125. - `--swagger <file>` analyze an OpenAPI/Swagger spec
  126. - `--prisma <file>` analyze a Prisma schema file
  127. - `--drizzle <path>` analyze Drizzle table definitions
  128. - `--mongoose <path>` analyze Mongoose schema/model files
  129. - `--no-install` skip auto-registration into AI clients
  130. - `--clients <list>` clients to register: `codex`, `claude-code`, `claude-desktop`, `vscode`, or `all` (default: `all`)
  131. ### `interactive`
  132. Prompts for which analyzers to run and which source files to include.
  133. ```bash
  134. npx mcpify-cli interactive
  135. ```
  136. ### `frontend [path]`
  137. Extracts UI actions only and can print raw JSON.
  138. ```bash
  139. npx mcpify-cli frontend ./examples/internal-tool --json
  140. ```
  141. ### `swagger <file>`
  142. Converts an OpenAPI or Swagger spec directly into MCP tools.
  143. ```bash
  144. npx mcpify-cli swagger ./tests/fixtures/swagger/petstore.yaml
  145. ```
  146. ### `audit [path]`
  147. Static safety audit over the discovered tools and workflows — no files written.
  148. ```bash
  149. npx mcpify-cli audit ./examples/express-api
  150. ```
  151. ### `simulate [path]`
  152. When `ANTHROPIC_API_KEY` is set, runs an AI simulation battery against the compiled tool surface.
  153. ```bash
  154. npx mcpify-cli simulate ./examples/express-api
  155. ```
  156. ## Architecture
  157. Built for the way agents actually operate. A layered system that keeps your application untouched while exposing exactly what agents need:
  158. 1. **AI Agents:** Any agent runtime (Claude, GPT, Custom) connects over MCP to start operating your software like a power user.
  159. 2. **MCP Layer:** Generated tools, resources, and prompts that map 1:1 to real surface in your app. The contract agents speak.
  160. 3. **Permissions:** Every call passes through scopes, audit logs, and rate limits before it touches your system.
  161. 4. **Workflows:** Discovered user journeys exposed as composable, stateful operations agents can chain.
  162. 5. **Your App:** Your existing stack (Frontend, Backend, Database, APIs) — untouched. MCPify reads it; it never rewrites it.
  163. ## Repository Structure
  164. ```txt
  165. mcpify/
  166. apps/
  167. docs/
  168. examples/
  169. landingPage/
  170. packages/
  171. ai-enhancer/
  172. backend-analyzer/
  173. cli/
  174. event-analyzer/
  175. frontend-analyzer/
  176. graph-engine/
  177. mcp-generator/
  178. monitoring/
  179. permissions/
  180. schema-engine/
  181. security/
  182. sync-engine/
  183. workflow-engine/
  184. tests/
  185. ```
  186. ## License
  187. See [LICENSE](LICENSE).

Ratings and comments

Sign in to rate this skill.

Sign in

Free to use. If it’s helped, pass it on.