Skip to content
Tutorials Center
My learning

Topics

Lesson 7 of 39 ยท Routes and Nested Layouts

Dynamic Segments, Optional Segments and Splat Routes

Learn dynamic URL params, optional segments, splat routes and 404 catch-all routes in React Router v7 framework mode, and how to validate params safely, with runnable stand-ins.

Most real sites have URLs that contain data: /products/42, /blog/hello-remix, /docs/guides/setup/install. You cannot write one route per product, so you describe the shape of the URL with a pattern, and the router hands the variable parts to your code as params. This lesson covers dynamic segments (:id), optional segments (:lang?), splats (*) and the catch-all 404 route, plus how to treat params as untrusted input.

Dynamic segments

A segment starting with : matches any single piece of the path (never one containing a slash).

TypeScript
// app/routes.ts (real, read-only)
route("products/:productId", "routes/product.tsx"),
route("teams/:teamId/members/:memberId", "routes/member.tsx"),

Your loader and component receive the values by name:

TSX
// app/routes/product.tsx (real, read-only)
import type { Route } from "./+types/product";

export async function loader({ params }: Route.LoaderArgs) {
  // params.productId is a string, typed from the route pattern
  const product = await getProduct(params.productId);
  return { product };
}

export default function Product({ loaderData, params }: Route.ComponentProps) {
  return <h1>{loaderData.product.name} (#{params.productId})</h1>;
}

The generated types read the pattern, so params.productId exists and params.nope is a compile error. Params are always strings and are already URL-decoded (/tags/c%2B%2B gives "c++").

Stand-in: a pattern matcher with params

This stand-in matches URL paths against patterns and returns the params, like the router does internally. Try changing the patterns or the URLs.

JS Stand-in: extract params from a URL

Press Run (or Ctrl + Enter) to see the console output here.

No output.

The second line returns null because dynamic segments never swallow a slash, which is also why /docs with an empty splat matches { "*": "" }.

Optional segments

Put a ? after a segment to make it optional. A single route then serves both /docs and /en/docs:

TypeScript
route(":lang?/docs", "routes/docs.tsx"),
route("categories/:category?", "routes/categories.tsx"),

In the loader, params.lang is string | undefined. Optional static segments work the same way (route("users/me?", ...)), but do not overdo it: a URL space that accepts many shapes is harder to reason about.

Splat routes

A trailing * matches the rest of the path, slashes included. The matched text lands in params["*"].

TSX
// route("files/*", "routes/files.tsx")  (real, read-only)
export async function loader({ params }: Route.LoaderArgs) {
  const path = params["*"]; // "photos/2025/trip.jpg" for /files/photos/2025/trip.jpg
  return { path, parts: path.split("/") };
}

Splats are right for file browsers, wiki-style pages, and docs trees whose depth you do not control.

The catch-all 404 route

A route with the path * at the end of the config matches everything else. Use it to render a branded "Page not found" page. Remember to set the HTTP status, otherwise search engines see a 200 for a page that does not exist (a "soft 404").

TSX
// route("*", "routes/not-found.tsx")  (real, read-only)
import { data } from "react-router";

export function loader() {
  throw data("Not Found", { status: 404 });
}

export default function NotFound() {
  return <h1>Page not found</h1>; // not rendered; the ErrorBoundary shows below
}

export function ErrorBoundary() {
  return <h1>Page not found</h1>;
}

Throwing data(..., { status: 404 }) sends a real 404 and renders the nearest error boundary. Throwing responses and 404s covers this in the error-handling chapter. When no route at all matches, the root error boundary shows a 404 automatically, so a catch-all route is mostly for a nicer, layout-aware page.

Specificity and conflicts

Given /products/new, /products/:id and /products/*, the static pattern wins, then the param, then the splat. See the previous lesson for the ranking stand-in.

Treat params as untrusted input

Anyone can type any URL. A param is just text that arrived from the network.

  • Convert and check: Number(params.id) can be NaN; Number.isInteger is not enough on its own if you need positive IDs.
  • Do not build file paths or SQL from params directly. Use parameterised queries and reject .. in file names.
  • Respond with a 404 when nothing matches or when the value is malformed, not a crash.
JS Stand-in: validate a param and answer 404 or 400

Press Run (or Ctrl + Enter) to see the console output here.

No output.

A malformed value gets 400, a well-formed value that does not exist gets 404, and only a real record gets a body. That distinction helps clients, logs and crawlers.

Linking to dynamic routes

Build links with the same pieces. Encode any user-provided value:

TSX
<Link to={`/tags/${encodeURIComponent(tag)}`}>{tag}</Link>
<Link to={{ pathname: "/products", search: "?sort=price" }}>Cheapest</Link>

Forgetting encodeURIComponent breaks tags like c++, a/b or 50%.

Check your understanding

Quick check ยท 4 questions

of 4 answered You scored / 4 ยท perfect!

  1. 1. What is the type of params.productId for the route products/:productId?

    Params are always strings, already URL-decoded. Convert and validate them yourself.
  2. 2. Where does a splat route put the matched rest of the path?

    The star segment is available as params['*'] and can contain slashes.
  3. 3. What does route(':lang?/docs', ...) match?

    The question mark makes the first segment optional, so params.lang may be undefined.
  4. 4. Why should a catch-all 404 page send a 404 status?

    A friendly page with status 200 is a soft 404 that crawlers may index. Throw data(..., { status: 404 }).

Summary

  • :name matches one segment, :name? makes it optional, * matches the rest of the path into params["*"].
  • Params are strings; validate and convert them, and respond with 400 or 404 for bad input.
  • Static beats dynamic beats splat; a final * route can render a custom not-found page.
  • Encode user text when building links.

Next: The Flat Routes Convention