Make your Next.js app agent-ready
Route handlers → MCP tools, typed clients and OpenAPI
Documentation    Quick Start    Performance

---

Vovk.ts

Vovk.ts makes your Next.js app agent-ready: every procedure is also an MCP tool, a typed client method and an OpenAPI operation. See the guide.

Vovk.ts adds an API layer on top of Next.js App Router Route Handlers. Its unit is the procedure: a typed function with its schema. From one procedure, Vovk.ts derives the HTTP endpoint, the local .fn() call, the type-safe client, the OpenAPI document and the AI tool with execute. You don't write a separate contract or glue code.

Requirements: Node.js 24.15+ (or Bun 1.4+, Deno 2.9+), Next.js 15+ and TypeScript 5.5+

Install to existing Next.js project

npx vovk-cli@latest init

See: https://vovk.dev/quick-install

Features

• 🧩 Plain Next.js: its routing, streaming, proxy.js and auth patterns, and deployment targets work as usual
• 🏗️ NestJS-style Controller → Service → Repository layers on top of Route Handlers, without the DI container or a custom server
• 📝 No separate contract: the schema comes from your controller code, so you don't maintain it by hand
• 🤖 AI tools from your API: controllers and generated RPC modules can become AI tools with an input schema and execute
• ⚡ Segments: split the API into parts, each with its own config and its own serverless function
• ✅ Typed requests with procedure(...): { params, query, body }
• 🔗 Third-party OpenAPI schemas as modules of the same client and tools (OpenAPI mixins)

What it looks like

A procedure is a typed, validated function. Define its inputs and output with procedure, and call it on the server for SSR/PPR, server actions or AI tool calls:

export default class UserController {
  static getUser = procedure({
    params: z.object({ id: z.string().uuid() }),
  }).handle(async (req, { id }) => {
    return UserService.getUserById(id);
  });
}

const user = await UserController.getUser.fn({ params: { id: '123e4567-e89b-12d3-a456-426614174000' } });

Services hold the business logic. Plain classes, no decorators; their types come from the procedure:

import type { VovkParams } from 'vovk';
import type UserController from './user-controller';

export default class UserService {
  static async getUserById(id: VovkParams<typeof UserController.getUser>['id']) {
    // ...
  }
}

Add an HTTP decorator, and the same procedure is also a Next.js Route Handler. The CLI generates a fetch-based client with the .fn() signature:

export default class UserController {
  @get('{id}')
  static getUser = procedure({
    params: z.object({ id: z.string().uuid() }),
  }).handle(async (req, { id }) => {
    return UserService.getUserById(id);
  });
}

import { UserRPC, PetstoreAPI } from '@/client';

const user = await UserRPC.getUser({ params: { id: '123e4567-e89b-12d3-a456-426614174000' } });
const pet = await PetstoreAPI.getPetById({ params: { petId: 1 } });

Add @operation, and the procedure is also an LLM tool. Pass controllers (in-process) or RPC modules (over HTTP) to deriveTools:

const tools = deriveTools({ modules: { UserRPC, TaskController, PetstoreAPI } });
console.log(tools); // [{ name, description, inputSchema, execute }, ...]

What one procedure becomes

From one function and its schema, Vovk.ts derives:

• the Next.js Route Handler: add an HTTP decorator to serve the procedure as an endpoint
• the local .fn() call, with the same call shape as the RPC client, for SSR, server components, server actions and AI tool calls
• the typed RPC client module, generated from the emitted schema, using fetch
• the OpenAPI 3.x document, from the same schema, so you don't maintain a separate spec
• the LLM tool with name, description, inputSchema and execute, from deriveTools
• a generated README.md that documents the client library

Agent Skills

15 topic skills teach a coding agent how to use Vovk.ts: Claude Code, Cursor, Windsurf, GitHub Copilot, Cline and the other agents that read skills. A skill loads only when it's relevant: "scaffold a new tenant" loads the multitenant skill, and "stream chat tokens" loads JSON Lines.

Install in your project:

npx skills add finom/vovk

In Claude Code, they also install as a plugin, with skill names that start with vovk::

/plugin marketplace add finom/vovk
/plugin install vovk@vovk

See: https://vovk.dev/about#agent-skills

Repository

• packages/: the npm packages vovk, vovk-cli, vovk-ajv, vovk-python and vovk-rust
• docs/: the vovk.dev site
• examples/: hello-world (hello-world.vovk.dev), kitchen-sink (examples.vovk.dev), realtime-kanban (kanban.vovk.dev) and multitenant (multitenant.vovk.dev)
• perf/: the overhead benchmarks
• test/: the integration test app

Links

• Docs: https://vovk.dev
• Quick Start: https://vovk.dev/quick-install
• Manual install: https://vovk.dev/manual-install
• Agent Skills: https://vovk.dev/about#agent-skills
• OpenAPI Mixins: https://vovk.dev/mixins
• Performance: https://vovk.dev/performance
• “Hello World” example app: https://github.com/finom/vovk/tree/main/examples/hello-world

---

License: MIT (see LICENSE).