A fork from next-safe-route that uses zod for schema validation.
next-zod-route
is a utility library for Next.js that provides type-safety and schema validation for Route Handlers/API Routes.
- ✅ Schema Validation: Automatically validates request parameters, query strings, and body content with built-in error handling.
- 🧷 Type-Safe: Works with full TypeScript type safety for parameters, query strings, and body content.
- 😌 Easy to Use: Simple and intuitive API that makes defining route handlers a breeze.
- 🔄 Flexible Response Handling: Return Response objects directly or return plain objects that are automatically converted to JSON responses.
- 🧪 Fully Tested: Extensive test suite to ensure everything works reliably.
- 🔐 Enhanced Middleware System: Powerful middleware system with pre/post handler execution, response modification, and context chaining.
- 🎯 Metadata Support: Add and validate metadata for your routes with full type safety.
- 🛡️ Custom Error Handling: Flexible error handling with custom error handlers for both middleware and route handlers.
npm install next-zod-route zod
Or using your preferred package manager:
pnpm add next-zod-route zod
yarn add next-zod-route zod
// app/api/hello/route.ts
import { createZodRoute } from 'next-zod-route';
import { z } from 'zod';
const paramsSchema = z.object({
id: z.string(),
});
const querySchema = z.object({
search: z.string().optional(),
});
const bodySchema = z.object({
field: z.string(),
});
const metadataSchema = z.object({
permission: z.string(),
role: z.enum(['admin', 'user']),
});
export const GET = createZodRoute()
.params(paramsSchema)
.query(querySchema)
.defineMetadata(metadataSchema)
.handler((request, context) => {
const { id } = context.params;
const { search } = context.query;
const { permission, role } = context.metadata!;
return Response.json({ id, search, permission, role }), { status: 200 };
});
export const POST = createZodRoute()
.params(paramsSchema)
.query(querySchema)
.body(bodySchema)
.handler((request, context) => {
// Next.js 15 use promise, but with .params we already unwrap the promise for you
const { id } = context.params;
const { search } = context.query;
const { field } = context.body;
return Response.json({ id, search, field }), { status: 200 };
});
To define a route handler in Next.js:
- Import
createZodRoute
andzod
. - Define validation schemas for params, query, body, and metadata as needed.
- Use
createZodRoute()
to create a route handler, chainingparams
,query
,body
, anddefineMetadata
methods. - Implement your handler function, accessing validated and type-safe params, query, body, and metadata through
context
.
next-zod-route
supports multiple request body formats out of the box:
- JSON: Automatically parses and validates JSON bodies.
- URL Encoded: Supports
application/x-www-form-urlencoded
data. - Multipart Form Data: Supports
multipart/form-data
, enabling file uploads and complex form data parsing.
The library automatically detects the content type and parses the body accordingly. For GET and DELETE requests, body parsing is skipped.
You can return responses in two ways:
- Return a Response object directly:
return Response.json({ data: 'value' }, { status: 200 });
- Return a plain object that will be automatically converted to a JSON response with status 200:
return { data: 'value' };
Metadata enable you to add static parameters to the route, for example to give permissions list to our application.
One powerful use case for metadata is defining required permissions for routes and checking them in middleware. This allows you to:
- Declare permissions statically at the route level
- Enforce permissions consistently across your application
- Keep authorization logic separate from your route handlers
Here's how to implement permission-based authorization:
// Define a schema for permissions metadata
const permissionsMetadataSchema = z.object({
requiredPermissions: z.array(z.string()).optional(),
});
// Create a middleware that checks permissions
const permissionCheckMiddleware = async ({ next, metadata, request }) => {
// Get user permissions from auth header, token, or session
const userPermissions = getUserPermissions(request);
// If no required permissions in metadata, allow access
if (!metadata?.requiredPermissions || metadata.requiredPermissions.length === 0) {
return next({ context: { authorized: true } });
}
// Check if user has all required permissions
const hasAllPermissions = metadata.requiredPermissions.every((permission) => userPermissions.includes(permission));
if (!hasAllPermissions) {
// Short-circuit with 403 Forbidden response
return new Response(
JSON.stringify({
error: 'Forbidden',
message: 'You do not have the required permissions',
}),
{
status: 403,
headers: { 'Content-Type': 'application/json' },
},
);
}
// Continue with authorized context
return next({ context: { authorized: true } });
};
// Use in your route handlers
export const GET = createZodRoute()
.defineMetadata(permissionsMetadataSchema)
.use(permissionCheckMiddleware)
.metadata({ requiredPermissions: ['read:users'] })
.handler((request, context) => {
// Only executed if user has 'read:users' permission
return Response.json({ data: 'Protected data' });
});
export const POST = createZodRoute()
.defineMetadata(permissionsMetadataSchema)
.use(permissionCheckMiddleware)
.metadata({ requiredPermissions: ['write:users'] })
.handler((request, context) => {
// Only executed if user has 'write:users' permission
return Response.json({ success: true });
});
export const DELETE = createZodRoute()
.defineMetadata(permissionsMetadataSchema)
.use(permissionCheckMiddleware)
.metadata({ requiredPermissions: ['admin:users'] })
.handler((request, context) => {
// Only executed if user has 'admin:users' permission
return Response.json({ success: true });
});
This pattern allows you to:
- Clearly document required permissions for each route
- Apply consistent authorization logic across your application
- Skip permission checks for public routes by not specifying required permissions
- Combine with other middleware for comprehensive request processing
You can add middleware to your route handler with the use
method. Middleware functions can add data to the context that will be available in your handler.
const loggingMiddleware = async ({ next }) => {
console.log('Before handler');
const startTime = performance.now();
const response = await next();
const endTime = performance.now() - start;
console.log(`After handler - took ${Math.round(endTime - startTime)}ms`);
return response;
};
const authMiddleware = async ({ request, metadata, next }) => {
try {
// Get the token from the request headers
const token = request.headers.get('authorization')?.split(' ')[1];
// You can access metadata in middleware
if (metadata?.role !== 'admin') {
throw new Error('Unauthorized');
}
// Validate the token and get the user
const user = await validateToken(token);
// Add context & continue chain
const response = await next({
context: { user },
});
// You can modify the response after the handler
return new Response(response.body, {
status: response.status,
headers: {
...Object.fromEntries(response.headers.entries()),
'X-User-Id': user.id,
},
});
} catch (error) {
// Errors in middleware are caught and handled by the error handler
throw error;
}
};
const permissionsMiddleware = async ({ metadata, next }) => {
// Metadata are optional and type-safe
const response = await next({
context: { permissions: metadata?.permissions ?? ['read'] },
});
return response;
};
export const GET = createZodRoute()
.defineMetadata(
z.object({
role: z.enum(['admin', 'user']),
permissions: z.array(z.string()).optional(),
}),
)
.use(loggingMiddleware)
.use(authMiddleware)
.use(permissionsMiddleware)
.handler((request, context) => {
// Access middleware data from context.data
const { user, permissions } = context.data;
// Access metadata from context.metadata
const { role } = context.metadata!;
return Response.json({ user, permissions, role });
});
Middleware functions receive:
request
: The request objectcontext
: The context object with data from previous middlewaresmetadata
: The validated metadata object (optional)next
: Function to continue the chain and add context
The middleware can:
- Execute code before/after the handler
- Modify the response
- Add context data through the chain
- Short-circuit the chain by returning a Response
- Throw errors that will be caught by the error handler
const timingMiddleware = async ({ next }) => {
console.log('Starting request...');
const start = performance.now();
const response = await next();
const duration = performance.now() - start;
console.log(`Request took ${duration}ms`);
return response;
};
const headerMiddleware = async ({ next }) => {
const response = await next();
return new Response(response.body, {
status: response.status,
headers: {
...Object.fromEntries(response.headers.entries()),
'X-Custom': 'value',
},
});
};
const middleware1 = async ({ next }) => {
const response = await next({
context: { value1: 'first' },
});
return response;
};
const middleware2 = async ({ context, next }) => {
// Access previous context
console.log(context.value1); // 'first'
const response = await next({
context: { value2: 'second' },
});
return response;
};
const authMiddleware = async ({ next }) => {
const isAuthed = false;
if (!isAuthed) {
return new Response(JSON.stringify({ error: 'Unauthorized' }), {
status: 401,
headers: { 'Content-Type': 'application/json' },
});
}
return next();
};
If you're upgrading from v0.1.x to v0.2.0, there are some changes to the middleware system:
const authMiddleware = async () => {
return { user: { id: 'user-123' } };
};
const route = createZodRoute()
.use(authMiddleware)
.handler((req, ctx) => {
const { user } = ctx.data;
return { data: user.id };
});
const authMiddleware = async ({ next }) => {
// Execute code before handler
console.log('Checking auth...');
// Add context & continue chain
const response = await next({
context: { user: { id: 'user-123' } },
});
// Modify response or execute code after
return new Response(response.body, {
headers: {
...Object.fromEntries(response.headers.entries()),
'X-User-Id': 'user-123',
},
});
};
const route = createZodRoute()
.use(authMiddleware)
.handler((req, ctx) => {
const { user } = ctx.data;
return { data: user.id };
});
Key changes in v0.2.0:
- Middleware must now accept an object with
request
,context
,metadata
, andnext
- Context is passed explicitly via
next({ context: {...} })
- Middleware can execute code before and after the handler
- Middleware can modify the response
- Middleware can short-circuit by returning a Response
- Error handling in middleware is now consistent with handler error handling
You can specify a custom error handler function to handle errors thrown in your route handler or middleware:
import { createZodRoute } from 'next-zod-route';
// Create a custom error class
class CustomError extends Error {
constructor(
message: string,
public status: number = 400,
) {
super(message);
this.name = 'CustomError';
}
}
// Create a route with a custom error handler
const safeRoute = createZodRoute({
handleServerError: (error: Error) => {
if (error instanceof CustomError) {
return new Response(JSON.stringify({ message: error.message }), { status: error.status });
}
// Default error response
return new Response(JSON.stringify({ message: 'Internal server error' }), { status: 500 });
},
});
export const GET = safeRoute
.use(async () => {
// This error will be caught by the custom error handler
throw new CustomError('Middleware error', 400);
})
.handler((request, context) => {
// This error will also be caught by the custom error handler
throw new CustomError('Handler error', 400);
});
By default, if no custom error handler is provided, the library will return a generic "Internal server error" message with a 500 status code to avoid information leakage.
When validation fails, the library returns appropriate error responses:
- Invalid params:
{ message: 'Invalid params' }
with status 400 - Invalid query:
{ message: 'Invalid query' }
with status 400 - Invalid body:
{ message: 'Invalid body' }
with status 400
Tests are written using Vitest. To run the tests, use the following command:
pnpm test
Contributions are welcome! For major changes, please open an issue first to discuss what you would like to change.
This project is licensed under the MIT License - see the LICENSE file for details.