--- title: "SvelteKit: agent-ready recipes | Agent Ready Checklist" description: "Every agent-ready recipe for SvelteKit: 31 rows of the checklist, the version read, the date and the vendor documentation each step came from." canonical: https://agentreadychecklist.com/platforms/sveltekit site: "Agent Ready Checklist" last_updated: 2026-11-05 section: "Platforms" --- # SvelteKit: agent-ready recipes | Agent Ready Checklist > Every agent-ready recipe for SvelteKit: 31 rows of the checklist, the version read, the date and the vendor documentation each step came from. Source: https://agentreadychecklist.com/platforms/sveltekit Site: Agent Ready Checklist Last reviewed: 2026-11-05 Agent endpoints: https://agentreadychecklist.com/llms.txt (index), https://agentreadychecklist.com/openapi.json (API), https://agentreadychecklist.com/mcp (MCP server), https://agentreadychecklist.com/ask (search) Access: https://agentreadychecklist.com/auth.md (no key). Pricing: https://agentreadychecklist.com/pricing.md (free). # SvelteKit: agent-ready recipes SvelteKit: 31 recipes against the 31 checklist rows, read on 2026-11-04. SvelteKit: 29 are direct, 2 are workarounds. SvelteKit recipes cite the page they were derived from, or the specification when the step is generic. A SvelteKit workaround names the real route: a proxy, an edge function or a domain you control. Version read for SvelteKit: SvelteKit. ## Version and documentation read Version SvelteKit (npm latest @sveltejs/kit 3.1.1, read 2026-11-04); docs at svelte.dev/docs/kit Read on 2026-11-04 Recipes 29 direct, 2 workaround, 0 not possible on the platform robots.txt No first-party doc: SvelteKit has no robots helper. Put robots.txt in static/ or write src/routes/robots.txt/+server.ts. Sitemap [svelte.dev/docs/kit/seo](https://svelte.dev/docs/kit/seo). SEO page shows a src/routes/sitemap.xml/+server endpoint returning XML. Structured data [svelte.dev/docs/kit/seo](https://svelte.dev/docs/kit/seo). Return SEO data from load and render it in in the layout; JSON-LD goes in the same head block. Status codes [svelte.dev/docs/kit/errors](https://svelte.dev/docs/kit/errors). error(404, "Not found") from @sveltejs/kit sets the status and renders +error.svelte. Headers and redirects [svelte.dev/docs/kit/load](https://svelte.dev/docs/kit/load). setHeaders in load functions; redirect() from @sveltejs/kit. Content negotiation [svelte.dev/docs/kit/hooks](https://svelte.dev/docs/kit/hooks). handle in src/hooks.server.ts sees every request and can return a different response by Accept header. llms.txt [svelte.dev/docs/kit/routing](https://svelte.dev/docs/kit/routing). No built-in; a +server.ts endpoint at src/routes/llms.txt/+server.ts. OpenAPI [svelte.dev/docs/kit/project-structure](https://svelte.dev/docs/kit/project-structure). No OpenAPI support; files in static/ are served as is. [The checklist table for SvelteKit](https://agentreadychecklist.com/platform/sveltekit/checklist/) shows every row with its first step. ## Verify with Ora SvelteKit has a direct recipe for 29 rows, and Ora's catalogue (read 2026-11-05) names a matching check for 25 of them, among them AR-FIND-01, AR-FIND-03, AR-FIND-04. After the steps are live, one audit shows which of those fixes the score picks up. ```bash npx ax@0.10 audit https://example.com --min-score 70 ``` One row that Ora checks is a workaround or not possible on SvelteKit: [AR-READ-05](https://agentreadychecklist.com/checklist/ar-read-05/) (code-fence-validity, workaround). Ora scores the site as it stands, so a row the platform cannot satisfy stays open until the workaround is in place. Run it again with `--force` after a deploy, so the audit skips Ora's cache. [Gate deploys on it in CI](https://agentreadychecklist.com/guides/gate-deploys-on-agent-readiness-in-ci/), [why this site recommends Ora](https://agentreadychecklist.com/verify-with-ora/), [the crosswalk of rows and checks](https://agentreadychecklist.com/checklist/ora-checks/). ## Recipes, row by row ### [AR-FIND-01](https://agentreadychecklist.com/checklist/ar-find-01/): Do not block agentsrecipe Write static/robots.txt with "User-agent: \*" and "Allow: /". Check the host WAF separately. Read 2026-11-04. [svelte.dev/docs/kit/project-structure](https://svelte.dev/docs/kit/project-structure) ### [AR-FIND-02](https://agentreadychecklist.com/checklist/ar-find-02/): Answers on one first-party pagerecipe Publish each answer once, as a page on your own domain. Put the fact (price, limit, version) in the page text, give the page one stable address, and link it from the home page and from llms.txt. Keep the page current: a second, older copy elsewhere on the site is the usual way an agent cites the wrong figure. Generic step from the AgentReady specification; not platform-specific. [github.com/agentready-org/standard/blob/main/spec/spec.md](https://github.com/agentready-org/standard/blob/main/spec/spec.md) ### [AR-FIND-03](https://agentreadychecklist.com/checklist/ar-find-03/): Sitemap listed in robots.txtrecipe Create src/routes/sitemap.xml/+server.ts: export async function GET() { return new Response(xml, { headers: { "Content-Type": "application/xml" } }); }, add export const prerender = true, and add a Sitemap line to robots.txt. Read 2026-11-04. [svelte.dev/docs/kit/seo](https://svelte.dev/docs/kit/seo) ### [AR-FIND-04](https://agentreadychecklist.com/checklist/ar-find-04/): AI Catalog at .well-knownrecipe Put static/.well-known/ai-catalog.json, or a prerendered src/routes/.well-known/ai-catalog.json/+server.ts returning json(catalog). Read 2026-11-04. [svelte.dev/docs/kit/routing](https://svelte.dev/docs/kit/routing) ### [AR-READ-01](https://agentreadychecklist.com/checklist/ar-read-01/): Answer in the initial HTMLrecipe Keep ssr enabled (the default) and load data in +page.server.ts or +page.ts load. For static docs add export const prerender = true. Do not set ssr = false on answer pages. Read 2026-11-04. [svelte.dev/docs/kit/page-options](https://svelte.dev/docs/kit/page-options) ### [AR-READ-02](https://agentreadychecklist.com/checklist/ar-read-02/): Real HTTP status codesrecipe In load call error(404, "Not found") when data is missing. In handle return new Response("Too Many Requests", { status: 429, headers: { "Retry-After": "60" } }) when throttling. With adapter-static, set fallback only for SPAs; a fallback page answers missing paths with 200 on many hosts. Read 2026-11-04. [svelte.dev/docs/kit/errors](https://svelte.dev/docs/kit/errors) ### [AR-READ-03](https://agentreadychecklist.com/checklist/ar-read-03/): Home page with real linksrecipe Use plain elements on the home page; SvelteKit renders them in the server HTML and upgrades navigation on the client. Read 2026-11-04. [svelte.dev/docs/kit/routing](https://svelte.dev/docs/kit/routing) ### [AR-READ-04](https://agentreadychecklist.com/checklist/ar-read-04/): Docs that answer agent tasksrecipe For each of your top five agent tasks, write a page that holds the steps as text, in order, with the exact names of the buttons, fields and settings. Steps that exist only in an image, a video or a script-built widget are not in the response a plain fetch receives. Fetch the page with curl and grep for the step. Generic step from the AgentReady specification; not platform-specific. [github.com/agentready-org/standard/blob/main/spec/spec.md](https://github.com/agentready-org/standard/blob/main/spec/spec.md) ### [AR-READ-05](https://agentreadychecklist.com/checklist/ar-read-05/): Fenced, language-tagged codeworkaround SvelteKit has no Markdown pipeline. Use mdsvex or render 
 with the language class.

Read 2026-11-04. [svelte.dev/docs/kit/project-structure](https://svelte.dev/docs/kit/project-structure)

### [AR-READ-06](https://agentreadychecklist.com/checklist/ar-read-06/): Link the discovery filesrecipe

In +layout.svelte add  and footer links to /llms.txt and /openapi.json.

Read 2026-11-04. [svelte.dev/docs/svelte/svelte-head](https://svelte.dev/docs/svelte/svelte-head)

### [AR-READ-07](https://agentreadychecklist.com/checklist/ar-read-07/): Ship llms.txtrecipe

Create src/routes/llms.txt/+server.ts returning text/plain and export const prerender = true, or drop static/llms.txt.

Read 2026-11-04. [svelte.dev/docs/kit/routing](https://svelte.dev/docs/kit/routing)

### [AR-READ-08](https://agentreadychecklist.com/checklist/ar-read-08/): JSON-LD on key pagesrecipe

In  render {@html ``} with data from load.

Read 2026-11-04. [svelte.dev/docs/kit/seo](https://svelte.dev/docs/kit/seo)

### [AR-READ-09](https://agentreadychecklist.com/checklist/ar-read-09/): Advertise a markdown mirrorrecipe

Add the alternate link in , or call setHeaders({ Link: "<…>; rel=\\"alternate\\"; type=\\"text/markdown\\"" }) in load.

Read 2026-11-04. [svelte.dev/docs/kit/load](https://svelte.dev/docs/kit/load)

### [AR-READ-11](https://agentreadychecklist.com/checklist/ar-read-11/): Pricing as first-party HTMLrecipe

Render src/routes/pricing/+page.svelte with prerender = true and the plans as HTML text.

Read 2026-11-04. [svelte.dev/docs/kit/page-options](https://svelte.dev/docs/kit/page-options)

### [AR-ACT-01](https://agentreadychecklist.com/checklist/ar-act-01/): Auth an agent can finishrecipe

Run an authorization server, or an identity provider, that issues scoped tokens, and serve its RFC 8414 metadata from the app as the CF-OAUTH-DISCOVERY recipe shows (/platforms/sveltekit/#cf-oauth-discovery). Also offer self-serve API keys with documented scopes.

Generic step from the AgentReady specification; not platform-specific. [github.com/agentready-org/standard/blob/main/spec/spec.md](https://github.com/agentready-org/standard/blob/main/spec/spec.md)

### [AR-ACT-02](https://agentreadychecklist.com/checklist/ar-act-02/): OpenAPI contractrecipe

Put openapi.json in static/ (served at /openapi.json) and link it.

Read 2026-11-04. [svelte.dev/docs/kit/project-structure](https://svelte.dev/docs/kit/project-structure)

### [AR-ACT-03](https://agentreadychecklist.com/checklist/ar-act-03/): MCP over Streamable HTTPrecipe

Add src/routes/mcp/+server.ts that exports POST and returns the MCP server over Streamable HTTP (a POST answered with JSON or an event stream). Keep a docs MCP public and gate a product MCP with OAuth and PKCE.

Derived from SvelteKit's documentation read on 2026-11-04. [svelte.dev/docs/kit/routing](https://svelte.dev/docs/kit/routing) and [github.com/agentready-org/standard/blob/main/spec/spec.md](https://github.com/agentready-org/standard/blob/main/spec/spec.md)

### [AR-ACT-04](https://agentreadychecklist.com/checklist/ar-act-04/): MCP server cardrecipe

Place the card at /.well-known/mcp/server-card.json the same way the AR-FIND-04 recipe places the AI Catalog (/platforms/sveltekit/#ar-find-04), and give it a serverInfo.name that matches your MCP server. Publish it only if you run an MCP server.

Generic step from the AgentReady specification; not platform-specific. [github.com/agentready-org/standard/blob/main/spec/spec.md](https://github.com/agentready-org/standard/blob/main/spec/spec.md)

### [AR-ACT-05](https://agentreadychecklist.com/checklist/ar-act-05/): MCP Apps viewrecipe

An MCP Apps view is a ui:// resource listed by resources/list on your MCP server. Add it to the server built for AR-ACT-03 (/platforms/sveltekit/#ar-act-03); the platform adds no step of its own. Publish it only if the product has a UI to hand back.

Generic step from the AgentReady specification; not platform-specific. [github.com/agentready-org/standard/blob/main/spec/spec.md](https://github.com/agentready-org/standard/blob/main/spec/spec.md)

### [AR-ACT-06](https://agentreadychecklist.com/checklist/ar-act-06/): SDKs and a CLIrecipe

Publish the SDK and the CLI from the product's own repository to npm or PyPI; the website platform is not involved. Then show the install command and a first call on a page of your own site so that agents reading the site find them. Verify with curl -s https://registry.npmjs.org//latest.

Generic step from the AgentReady specification; not platform-specific. [github.com/agentready-org/standard/blob/main/spec/spec.md](https://github.com/agentready-org/standard/blob/main/spec/spec.md)

### [AR-ACT-07](https://agentreadychecklist.com/checklist/ar-act-07/): WebMCP toolsrecipe

In +page.svelte register tools in onMount (browser only): if (document.modelContext) document.modelContext.registerTool({…}).

Read 2026-11-04. [svelte.dev/docs/kit/page-options](https://svelte.dev/docs/kit/page-options)

### [AR-ACT-08](https://agentreadychecklist.com/checklist/ar-act-08/): Agent paymentsrecipe

Add src/routes/paid/+server.ts that exports GET and returns HTTP 402 with a payment payload on the paid path only (x402 or MPP), or a checkout handler. Put no CAPTCHA in front of a verified agent.

Derived from SvelteKit's documentation read on 2026-11-04. [svelte.dev/docs/kit/routing](https://svelte.dev/docs/kit/routing) and [github.com/agentready-org/standard/blob/main/spec/spec.md](https://github.com/agentready-org/standard/blob/main/spec/spec.md)

### [AR-ACT-09](https://agentreadychecklist.com/checklist/ar-act-09/): NLWeb /ask endpointrecipe

Add src/routes/ask/+server.ts that exports GET and returns schema.org-shaped JSON for the query parameter. Mirror the same ask tool over MCP.

Derived from SvelteKit's documentation read on 2026-11-04. [svelte.dev/docs/kit/routing](https://svelte.dev/docs/kit/routing) and [github.com/agentready-org/standard/blob/main/spec/spec.md](https://github.com/agentready-org/standard/blob/main/spec/spec.md)

### [AR-ACT-10](https://agentreadychecklist.com/checklist/ar-act-10/): An MCP server or a CLIrecipe

A CLI meets AR-ACT-10 on any platform: publish it as described in the AR-ACT-06 recipe (/platforms/sveltekit/#ar-act-06). An MCP server also meets it; how SvelteKit hosts one is in the AR-ACT-03 recipe (/platforms/sveltekit/#ar-act-03). Do not ship an empty surface to tick the box.

Generic step from the AgentReady specification; not platform-specific. [github.com/agentready-org/standard/blob/main/spec/spec.md](https://github.com/agentready-org/standard/blob/main/spec/spec.md)

### [CF-CONTENT-SIGNALS](https://agentreadychecklist.com/checklist/cf-content-signals/): Content-Signal linerecipe

Add a Content-Signal line under User-agent: \* in static/robots.txt.

Read 2026-11-04. [svelte.dev/docs/kit/project-structure](https://svelte.dev/docs/kit/project-structure)

### [CF-MARKDOWN-NEGOTIATION](https://agentreadychecklist.com/checklist/cf-markdown-negotiation/): Markdown by Acceptrecipe

In src/hooks.server.ts handle: if (event.request.headers.get("accept")?.includes("text/markdown")) return new Response(markdownFor(event.url.pathname), { headers }); otherwise return resolve(event).

Read 2026-11-04. [svelte.dev/docs/kit/hooks](https://svelte.dev/docs/kit/hooks)

### [CF-PAY-PER-CRAWL](https://agentreadychecklist.com/checklist/cf-pay-per-crawl/): Pay per crawl, HTTP 402workaround

Pay per crawl is a feature of Cloudflare AI Crawl Control, set per Cloudflare zone and in closed beta on 2026-11-04, so SvelteKit does not set it, and a crawler is answered 402 only on a zone that has it. The documentation read lists no built-in feature for this row. Workaround: put the domain on a Cloudflare zone, where SvelteKit allows a proxy or custom DNS in front, and set the price there. Test with curl -A "GPTBot/1.0".

Derived from Cloudflare's pay per crawl documentation read on 2026-11-04. [developers.cloudflare.com/ai-crawl-control/features/pay-per-crawl/what-is-pay-per-crawl/](https://developers.cloudflare.com/ai-crawl-control/features/pay-per-crawl/what-is-pay-per-crawl/) and [developers.cloudflare.com/ai-crawl-control/features/pay-per-crawl/what-is-pay-per-crawl/](https://developers.cloudflare.com/ai-crawl-control/features/pay-per-crawl/what-is-pay-per-crawl/)

### [CF-OAUTH-DISCOVERY](https://agentreadychecklist.com/checklist/cf-oauth-discovery/): OAuth discovery filesrecipe

Return the RFC 8414 metadata from src/routes/.well-known/oauth-authorization-server/+server.ts with json(metadata).

Read 2026-11-04. [svelte.dev/docs/kit/routing](https://svelte.dev/docs/kit/routing)

### [CF-AGENT-SKILLS](https://agentreadychecklist.com/checklist/cf-agent-skills/): Agent Skills indexrecipe

Place static/.well-known/agent-skills/index.json and the SKILL.md files.

Read 2026-11-04. [svelte.dev/docs/kit/project-structure](https://svelte.dev/docs/kit/project-structure)

### [VERCEL-NEGOTIATION-HEADERS](https://agentreadychecklist.com/checklist/vercel-negotiation-headers/): Markdown headersrecipe

Set headers { "Content-Type": "text/markdown; charset=utf-8", "Vary": "Accept" } on the markdown response and append Vary: Accept to the HTML response from resolve(event).

Read 2026-11-04. [svelte.dev/docs/kit/hooks](https://svelte.dev/docs/kit/hooks)

### [VERCEL-ROBOTS-EXCLUSIONS](https://agentreadychecklist.com/checklist/vercel-robots-exclusions/): Robots exclusionsrecipe

Open the robots.txt that SvelteKit serves for your site and read every Disallow; each one must have a reason you can state. How the file is produced here: No first-party doc: SvelteKit has no robots helper. Put robots.txt in static/ or write src/routes/robots.txt/+server.ts.

Derived from SvelteKit's documentation read on 2026-11-04. [svelte.dev/docs/kit/project-structure](https://svelte.dev/docs/kit/project-structure) and [vercel.com/kb/guide/make-your-site-readable-by-ai-agents](https://vercel.com/kb/guide/make-your-site-readable-by-ai-agents)

## Vendors on this platform

-   [SvelteKit](https://agentreadychecklist.com/vendors/sveltekit/)Framework, 2 of 11 checklist rows read met by default

## Sources

-   [ora.ai/api/checks, contract 1.38.0](https://ora.ai/api/checks?utm_source=agentreadychecklist.com&utm_medium=referral&utm_campaign=platform)read 2026-11-05
-   [registry.npmjs.org/ax, the ax README](https://registry.npmjs.org/ax)read 2026-11-05
-   [svelte.dev/docs/kit/seo](https://svelte.dev/docs/kit/seo)read 2026-11-04
-   [svelte.dev/docs/kit/errors](https://svelte.dev/docs/kit/errors)read 2026-11-04
-   [svelte.dev/docs/kit/load](https://svelte.dev/docs/kit/load)read 2026-11-04
-   [svelte.dev/docs/kit/hooks](https://svelte.dev/docs/kit/hooks)read 2026-11-04
-   [svelte.dev/docs/kit/routing](https://svelte.dev/docs/kit/routing)read 2026-11-04
-   [svelte.dev/docs/kit/project-structure](https://svelte.dev/docs/kit/project-structure)read 2026-11-04
-   [github.com/agentready-org/standard/blob/main/spec/spec.md](https://github.com/agentready-org/standard/blob/main/spec/spec.md)read 2026-11-04
-   [svelte.dev/docs/kit/page-options](https://svelte.dev/docs/kit/page-options)read 2026-11-04
-   [svelte.dev/docs/svelte/svelte-head](https://svelte.dev/docs/svelte/svelte-head)read 2026-11-04
-   [developers.cloudflare.com/ai-crawl-control/features/pay-per-crawl/what-is-pay-per-crawl/](https://developers.cloudflare.com/ai-crawl-control/features/pay-per-crawl/what-is-pay-per-crawl/)read 2026-11-04
-   [vercel.com/kb/guide/make-your-site-readable-by-ai-agents](https://vercel.com/kb/guide/make-your-site-readable-by-ai-agents)read 2026-11-04

Edition

Data

2026-11-04

Reviewed

2026-11-05

Spec

AgentReady v2.0, 273d2d7

Crawl

13,071 rows, 2026-11-04

Ora

132 checks, contract 1.38.0, 2026-11-05

Verify with Ora

The scanner this site recommends for the whole checklist. 132 checks, a score out of 100, no account.

`npx ax@0.10 audit  --min-score `

[Why Ora, and its limits](https://agentreadychecklist.com/verify-with-ora/)

Method

-   [How rows, commands and rates are made](https://agentreadychecklist.com/method/)
-   [How an Ora check is matched to a row](https://agentreadychecklist.com/checklist/ora-checks/)
-   [Pass rates by rank band](https://agentreadychecklist.com/data/by-rank/)

Sources and read dates sit at the foot of the page.

Last reviewed 2026-11-05. AgentReady v2.0, read 2026-11-04.