remix-v2-data-flow

$npx mdskill add existential-birds/beagle/remix-v2-data-flow

Load and mutate data in Remix v2 routes with typed APIs.

  • Fetch data in loaders and handle mutations in actions.
  • Uses Remix loaders, actions, json, defer, and Form.
  • Selects loader/action based on HTTP method and route params.
  • Returns typed data via useLoaderData and useActionData hooks.

SKILL.md

.github/skills/remix-v2-data-flowView on GitHub ↗
---
name: remix-v2-data-flow
description: Remix v2 data loading and mutations. Use when writing loaders, actions, deferred data, revalidation logic, or pending state. Triggers on loader, action, useLoaderData, useActionData, json(), defer(), <Await>, shouldRevalidate, useRevalidator, useNavigation, useTransition (v1 holdover).
---

# Remix v2 Data Flow

## Quick Reference

**Loader + typed read**:
```tsx
import { json, type LoaderFunctionArgs } from "@remix-run/node";
import { useLoaderData } from "@remix-run/react";

export async function loader({ request }: LoaderFunctionArgs) {
  const invoices = await db.invoice.findMany();
  return json({ invoices });
}

export default function Invoices() {
  // typeof loader is a type ANNOTATION (not assertion) — drives SerializeFrom<T>.
  const { invoices } = useLoaderData<typeof loader>();
  return <InvoiceList invoices={invoices} />;
}
```

**Action + redirect-after-success (PRG)**:
```tsx
import { json, redirect, type ActionFunctionArgs } from "@remix-run/node";
import { useActionData, Form } from "@remix-run/react";

export async function action({ request }: ActionFunctionArgs) {
  const form = await request.formData();
  const parsed = NewProject.safeParse(Object.fromEntries(form));
  if (!parsed.success) return json({ errors: parsed.error.flatten().fieldErrors }, { status: 400 });
  const project = await db.project.create({ data: parsed.data });
  return redirect(`/projects/${project.id}`);
}
```

## Canonical APIs

Route modules export `loader` / `action`; components read results via `useLoaderData<typeof loader>()` and `useActionData<typeof action>()`. After every action, Remix automatically revalidates the loaders of all matching routes on the page, so the UI stays consistent with the server without manual cache invalidation.

Signatures:

- `loader: ({ request, params, context }: LoaderFunctionArgs) => Response | Promise<Response>` — server-only read, runs on SSR and on client navigations via fetch.
- `action: ({ request, params, context }: ActionFunctionArgs) => Response | Promise<Response>` — server-only handler for non-GET requests (POST/PUT/PATCH/DELETE).
- `json(data, init?: number | ResponseInit): TypedResponse<typeof data>` — ergonomic JSON `Response` wrapper with status/headers.
- `redirect(url, init?: number | ResponseInit): TypedResponse<never>` — 30x response; default 302.

Imports: `@remix-run/node` for server utilities (`json`, `redirect`, `defer`, type args) on Node; substitute `@remix-run/cloudflare` or `@remix-run/deno` for those targets. Hooks and components come from `@remix-run/react`.

## Type Annotations, Not Assertions

`useLoaderData<typeof loader>()` is a **type annotation**, not a `as`-style assertion. The generic feeds `SerializeFrom<typeof loader>`, which models the wire-format transformation: `Date` becomes `string`, `Map`/`Set` collapse, `undefined` fields are stripped, class methods vanish. If you call `data.createdAt.getFullYear()` on a `Date` field, that's a runtime bug — the type already says `string`.

## When `json()` Is Optional in v2

v2 did **not** change the underlying contract: loaders and actions must return a `Response`. `json()` is the ergonomic wrapper that sets `application/json` and lets you supply status / headers. Bare object returns work in v2 (Remix auto-wraps as `json()`), but `json()` is preferred for explicit status, headers, and clean `TypedResponse<T>` typing. Reach for `json()` whenever you need:

- A non-200 status code (e.g. `{ status: 400 }` for validation errors).
- Custom headers (caching, `Set-Cookie`).
- An explicit `TypedResponse<T>` for clean `useLoaderData<typeof loader>()` inference.

## Throwing for Short-Circuits

Throwing a `Response` from a loader or action exits the data function immediately. Use this for auth guards (`throw redirect("/login")`) and 404s (`throw new Response("Not Found", { status: 404 })` or `throw json({ message }, { status: 404 })`). Throwing a plain `Error` will not be classified as a route response by `useRouteError()` / `isRouteErrorResponse()`.

## Streaming Rejections: `<Await errorElement>` + `useAsyncError()`

When a promise passed through `defer()` rejects, an `<Await errorElement={...}>` boundary catches it inline — without it, the rejection bubbles to the route's `ErrorBoundary` and tears down the whole page, defeating the streaming benefit. Inside the `errorElement`, call `useAsyncError()` (from `@remix-run/react`) to read the rejection value — this is the streaming analogue of `useRouteError()`.

```tsx
function ReviewsError() {
  const error = useAsyncError(); // typed as `unknown`
  return <p>Failed to load reviews: {String(error)}</p>;
}

<Suspense fallback={<ReviewsSkeleton />}>
  <Await resolve={reviews} errorElement={<ReviewsError />}>
    {(r) => <ReviewList reviews={r} />}
  </Await>
</Suspense>
```

## Sensitive Data

Everything returned from a loader travels to the browser as JSON. Project to a safe DTO (`{ id, email, name }`) before returning; never return the full Prisma `User`, password hashes, API keys, or internal flags. Loaders execute server-only — but the *return value* is shipped to the client wholesale.

## Mutations Belong in Actions

Loaders run on every GET navigation and may be invoked speculatively by prefetch; they also re-run during automatic revalidation. Anything that mutates persistent state must live in `action`, reached via `<Form method="post">` or `useFetcher`. Calling `fetch()` directly from a component to hit a Remix route bypasses revalidation, pending state, and progressive enhancement — use `useFetcher().submit()` / `useFetcher().load()` instead.

## Routing Gotchas to Remember

- **Only the deepest matching action runs.** Index routes nested under a layout collide unless you target them with `<Form action="/things?index" method="post" />`.
- **`useActionData` is scoped to the current route.** It cannot access action results from parent or child routes; to share, lift the action or use a `useFetcher` with `key`.
- **Headers from the leaf loader win.** Remix only uses the deepest matching `headers` export. Parent caching policies are ignored unless you explicitly merge `parentHeaders`.
- **Default revalidation revalidates ALL routes on the page after an action** — even those whose params didn't change. Use `shouldRevalidate` to opt out of expensive parent loaders.

## Pending State (v1 → v2)

`useTransition` is removed in v2 — use `useNavigation`. The `submission` object is flattened directly onto the navigation in v2 (and the fetcher likewise; both `nav.formData`/`nav.formMethod` and `fetcher.formData`/`fetcher.formMethod` are flat). `formMethod` is now **UPPERCASE** in v2 (`"POST"`, not `"post"`); comparisons like `nav.formMethod === "post"` silently never match. `fetcher.type` is also gone — branch on `fetcher.state` plus presence of `fetcher.formData`.

GET submissions go `idle → loading → idle`. POST flow goes `idle → submitting → loading → idle`. Spinners gated only on `"submitting"` will miss GET forms. GET submissions still populate `nav.formData` and `nav.formMethod === "GET"` during the `loading` phase, so filter-form pending UI should branch on `formData` presence, not on `state === 'submitting'`. For `useFetcher`, `submitting` applies to BOTH GET (`<fetcher.Form method='get'>` and `fetcher.submit(..., {method:'get'})`) and non-GET; only `fetcher.load()` skips `submitting`. This is the inverse of `useNavigation`, which skips `submitting` for GET.

## Gates (decision sequencing)

Answer **in order**. **Pass** means the condition is true; pick the API on the same line and **stop**.

### `loader` vs `useEffect`

1. **Is the data needed for correct first render** of this route (SSR, prefetch, automatic revalidation after actions)?
   - **Pass →** `loader` + `useLoaderData<typeof loader>()`. **Stop.**
   - **Fail →** Step 2.
2. **Is the fetch driven by post-mount user interaction, timer, or subscription** (not route entry)?
   - **Pass →** `useEffect` / event handlers. **Stop.**
   - **Fail →** Prefer loader + revalidation; do not mirror navigation inside an effect.

### `json()` vs raw `Response` vs `defer()`

1. **Do any returned fields need to stream** (slow query, expensive aggregation) while the page renders fast?
   - **Pass →** `defer({ critical: await…, slow: promiseWithoutAwait })` + `<Suspense><Await>…</Await></Suspense>`. **Stop.**
   - **Fail →** Step 2.
2. **Do you need a custom status code, custom headers, or explicit `TypedResponse<T>` typing**?
   - **Pass →** `json(data, init)` (or `redirect(url, init)` for 3xx). **Stop.**
   - **Fail →** Step 3.
3. **Do you need a non-JSON body** (binary, plain text, streamed file)?
   - **Pass →** Build a raw `new Response(body, init)`. **Stop.**
   - **Fail →** Default to `json(data)` — it's the documented v2 contract for object payloads.

### `<Form>` / route action vs `useFetcher`

1. **Should the URL or history stack change** (bookmark / share / back returns to prior screen)?
   - **Pass →** `<Form method="post">` posting to a route `action`. **Stop.**
   - **Fail →** Step 2.
2. **Mutation stays on the same route** (inline edit, list-row toggle, popover, optimistic UI)?
   - **Pass →** `useFetcher()` / `fetcher.Form` / `fetcher.submit()`. **Stop.**

## Additional Documentation

- **Loaders**: See [references/loaders.md](references/loaders.md) for `loader` signature, typed `useLoaderData`, `json()` vs raw `Response`, `redirect()`, throwing, sensitive-data filtering, params handling.
- **Actions**: See [references/actions.md](references/actions.md) for `action` signature, FormData parsing, `useActionData<typeof action>()`, zod/valibot validation, redirect-after-success.
- **Defer & Await**: See [references/defer-await.md](references/defer-await.md) for `defer()` + `<Await>` + `<Suspense>`, when streaming helps TTFB, error handling.
- **Revalidation & Pending State**: See [references/revalidation.md](references/revalidation.md) for automatic revalidation, `shouldRevalidate`, `useRevalidator`, `useNavigation` (plus v1 `useTransition` rename).

## v1 → v2 Quick Diff

| Concern              | v1                                  | v2                                               |
|----------------------|-------------------------------------|--------------------------------------------------|
| Navigation hook      | `useTransition()`                   | `useNavigation()`                                |
| Submission shape     | `transition.submission.formMethod`  | flat `nav.formMethod` / `nav.formData`           |
| `formMethod` casing  | `"post"`                            | `"POST"` (UPPERCASE)                             |
| Fetcher type field   | `fetcher.type === "actionSubmission"` | branch on `fetcher.state` + `fetcher.formData` |
| Loader args type     | `LoaderArgs` / `ActionArgs`         | `LoaderFunctionArgs` / `ActionFunctionArgs`      |
| Returning data       | `json(data)` required               | `json(data)` still the documented contract       |

More from existential-birds/beagle

SkillDescription
adr-decision-extractionUse when you need to mine a conversation, session transcript, or design discussion for architectural decisions before writing ADRs. Identifies problem-solution pairs, trade-off debates, technology choices, and explicit \"[ADR]\" tags. Triggers on \"what decisions did we make\", \"extract decisions from this chat\", \"find the choices in our discussion\", or \"summarize architectural decisions\". Also useful after long planning sessions to capture decisions that were made implicitly. Does NOT write ADR documents \u2014 use adr-writing or write-adr for that.
adr-writingUse when writing or formatting an ADR document using the MADR template, applying Definition of Done (E.C.A.D.R.) criteria, or verifying ADR completeness. Triggers on \"write the ADR\", \"format as MADR\", \"check ADR quality\", \"mark gaps in ADR\". Also triggers when a decision has been extracted and needs to become a document. Does NOT extract decisions from conversations (use adr-decision-extraction) or orchestrate the full extract-confirm-write workflow (use write-adr).
agent-architecture-analysisUse when auditing an agent codebase against the 12-Factor Agents methodology, reviewing LLM-powered system architecture, or assessing agentic app compliance. Triggers on \"analyze agent architecture\", \"12-factor audit\", \"how compliant is this agent\", or \"evaluate this LLM app\". Also applies when comparing frameworks or planning agent improvements. Not for quick checklists \u2014 this performs deep per-factor codebase analysis with file-level evidence.
app-intents-code-reviewReviews App Intents code for intent structure, entities, shortcuts, and parameters. Use when reviewing code with import AppIntents, @AppIntent, AppEntity, AppShortcutsProvider, or @Parameter.
artifact-analysisUse when the user wants a cited, structured read of local documents and project knowledge. Triggers on: \"analyze these docs\", \"scan my project for context\", \"read the docs folder\", \"summarize what's in .beagle/concepts/\", \"extract context from docs/\", \"what's in this folder\", \"go read everything in X and tell me what's there\". Also invoked programmatically by other beagle skills (prfaq-beagle Ignition, brainstorm-beagle reference points, strategy-interview context grounding) via the companion contract. Does NOT trigger on codebase lookups (\"find this function\", \"search the repo\"), web research (use web-research), LLM-as-judge evaluation (use llm-judge), or document editing (use humanize-beagle). Produces a written scan plan, parallel-subagent findings, and a cited synthesis report on disk — never inline prose, never unsourced claims.
axum-code-reviewReviews axum web framework code for routing patterns, extractor usage, middleware, state management, and error handling. Use when reviewing Rust code that uses axum, tower, or hyper for HTTP services. Covers axum 0.7+ patterns including State, Path, Query, Json extractors.
brainstorm-beagleUse when the user has a fuzzy idea and wants to shape it into a concrete project spec before planning or building. Triggers on: \"brainstorm this\", \"I have an idea for...\", \"help me think through this project\", \"what should I build\", \"spec this out\". Also catches vague feature descriptions needing structured questioning to clarify scope. Does NOT write code, plan implementation, review strategy docs, or run strategy interviews \u2014 produces a WHAT/WHY spec through dialogue, not a HOW plan.
bubbletea-code-reviewReviews BubbleTea TUI code for proper Elm architecture, model/update/view patterns, and Lipgloss styling. Use when reviewing terminal UI code using charmbracelet/bubbletea.
cloudkit-code-reviewReviews CloudKit code for container setup, record handling, subscriptions, and sharing patterns. Use when reviewing code with import CloudKit, CKContainer, CKRecord, CKShare, or CKSubscription.
combine-code-reviewReviews Combine framework code for memory leaks, operator misuse, and error handling. Use when reviewing code with import Combine, AnyPublisher, @Published, PassthroughSubject, or CurrentValueSubject.