Remove page-to-markdown capability

docs/contributing/environment-variables.md

@@ -91,23 +91,21 @@ Optional Worker secret:
   dev uses offline search while `WRANGLER_IS_LOCAL_DEV` is set or the binding is
   missing.
 
-## Cloudflare API (Worker + Browser Rendering + Email)
+## Cloudflare API (Worker + Email)
 
 Optional Worker secrets/vars (see `packages/worker/src/env-schema.ts` and
 `packages/worker/src/mcp/cloudflare/cloudflare-rest-client.ts`):
 
 - `CLOUDFLARE_API_TOKEN` — Cloudflare API token used by the internal API client
-  (`Authorization: Bearer ...`) for `page_to_markdown` Browser Rendering and the
-  Cloudflare Email sender. User Cloudflare API calls from codemode use saved
-  secrets and secret-aware `fetch` (see
-  `docs/contributing/skill-patterns/cloudflare-api-v4.md`). Local `npm run dev`
-  sets this to the Cloudflare mock token unless `AI_MODE=remote` or
-  `SKIP_CLOUDFLARE_MOCK=1`; when unset and no mock is attached, the billed
-  `page_to_markdown` Browser Rendering fallback fails fast with a setup hint.
-- `CLOUDFLARE_ACCOUNT_ID` — Cloudflare account id required by the
-  `page_to_markdown` capability when it falls back to Browser Rendering
-  `POST /client/v4/accounts/{account_id}/browser-rendering/markdown`. This is a
-  Worker var (not a secret) and should match the account behind
+  (`Authorization: Bearer ...`) for Worker-side Cloudflare REST calls such as
+  the Cloudflare Email sender. User
+  Cloudflare API calls from codemode use saved secrets and secret-aware `fetch`
+  (see `docs/contributing/skill-patterns/cloudflare-api-v4.md`). Local
+  `npm run dev` sets this to the Cloudflare mock token unless
+  `AI_MODE=remote` or `SKIP_CLOUDFLARE_MOCK=1`.
+- `CLOUDFLARE_ACCOUNT_ID` — Cloudflare account id required by the Cloudflare
+  Email Service REST API fallback used by local mocks and preview deploys. This
+  is a Worker var (not a secret) and should match the account behind
   `CLOUDFLARE_API_TOKEN`.
 - `CLOUDFLARE_API_BASE_URL` — API base URL; defaults to
   `https://api.cloudflare.com` when unset, including for outbound email sending.

docs/contributing/mock-api-servers.md

@@ -17,8 +17,8 @@ Worker package so it can also be deployed alongside the main app.
    and set `ACME_API_BASE_URL` to the mock Worker origin.
 
 See `packages/mock-servers/cloudflare/` for a small Cloudflare API v4 subset
-mock used by tests and `page_to_markdown` / internal API client (started by
-`npm run dev` unless `SKIP_CLOUDFLARE_MOCK=1`).
+mock used by tests and the internal API client (started by `npm run dev`
+unless `SKIP_CLOUDFLARE_MOCK=1`).
 
 ### Tips
 

docs/contributing/setup-manifest.md

@@ -72,9 +72,8 @@ automatically:
   Cloudflare AI Gateway)
 - `CLOUDFLARE_ACCOUNT_ID` (required for local development when `AI_MODE=remote`
   so Wrangler can authenticate Workers AI requests against the correct account;
-  also required when using the `page_to_markdown` capability's Cloudflare
-  Browser Rendering fallback against the live API and for the Cloudflare Email
-  Service REST API fallback used by local mocks and preview deploys)
+  also required for the Cloudflare Email Service REST API fallback used by
+  local mocks and preview deploys)
 - `CLOUDFLARE_API_TOKEN` (required for local development when `AI_MODE=remote`
   so Wrangler can authenticate Workers AI requests; also reused by the
   Cloudflare Email Service REST API fallback when local/preview email is routed
@@ -106,7 +105,8 @@ secrets from `packages/worker/.env`.
 Configure these GitHub Actions secrets and variables for workflows:
 
 - `CLOUDFLARE_API_TOKEN` (Workers deploy + D1 edit access on the correct
-  account)
+  account; also reused for remote AI and Cloudflare API workflows that run with
+  account secrets + skills)
 - `COOKIE_SECRET` (same format as local)
 - `APP_BASE_URL` (optional GitHub Actions **variable**, used by the production
   deploy as the canonical public app origin and written into the generated
@@ -136,7 +136,10 @@ How to get/set each value:
 
 - `CLOUDFLARE_API_TOKEN`
   - In Cloudflare Dashboard, create an API Token with permissions to deploy
-    Workers and edit D1 on the target account.
+    Workers and edit D1 on the target account. This is the same token to reuse
+    for remote AI and Cloudflare API workflows that run with account secrets +
+    skills; when you do, also include the product permissions needed for those
+    APIs.
   - In GitHub: `Settings` → `Secrets and variables` → `Actions` →
     `New repository secret`.
 - `COOKIE_SECRET`
@@ -174,13 +177,6 @@ How to get/set each value:
   - In GitHub: **Settings → Secrets and variables → Actions → Variables**, add
     `SENTRY_ORG` and `SENTRY_PROJECT` with your Sentry slugs (for example from
     `npx @sentry/wizard@latest -i sourcemaps`).
-- `CLOUDFLARE_API_TOKEN` (required for remote AI; used by the Worker for
-  `page_to_markdown` Browser Rendering; user API calls use account secrets +
-  skills)
-  - Create a Cloudflare API token with the account permissions needed for the
-    product APIs you want to call. This same secret already powers production
-    deploys and the `page_to_markdown` Browser Rendering path. For Browser
-    Rendering fallback, include the **Browser Rendering - Edit** permission.
 - `CAPABILITY_REINDEX_SECRET` (optional)
   - Generate a long random secret (for example `openssl rand -hex 32`), store it
     as the repository secret `CAPABILITY_REINDEX_SECRET`, and let the deploy

docs/contributing/setup.md

@@ -35,8 +35,8 @@ Quick notes for getting a local kody environment running.
 - `npm run dev` (starts mock API servers automatically, the main worker, and the
   local home connector; it sets `AI_MODE=mock`, `AI_MOCK_BASE_URL`, and
   `CLOUDFLARE_API_BASE_URL` + `CLOUDFLARE_API_TOKEN` + `CLOUDFLARE_ACCOUNT_ID`
-  to the local Cloudflare API mock Worker for `page_to_markdown`, the internal
-  Cloudflare API client, and local email sending. Unless you already set
+  to the local Cloudflare API mock Worker for the internal Cloudflare API client
+  and local email sending. Unless you already set
   `CLOUDFLARE_EMAIL_FROM`, the launcher also defaults it to `reset@kody.dev`.
   Set `SKIP_CLOUDFLARE_MOCK=1` to skip the local Cloudflare mock entirely. The
   home connector receives the resolved worker origin via `WORKER_BASE_URL`. When

docs/contributing/skill-patterns/cloudflare-api-v4.md

@@ -25,8 +25,6 @@ Capability **allowlists** on secrets apply to **capability inputs** that use
 
 All API paths must be under **`/client/v4/`** (see
 [Cloudflare API docs](https://developers.cloudflare.com/fundamentals/api/how-to/make-api-calls/)).
-The internal `CloudflareRestClient` used for Browser Rendering enforces the same
-shape.
 
 ## Example skill (save via `meta_save_skill`)
 
@@ -61,7 +59,5 @@ For **POST / PUT / PATCH / DELETE**, set `Content-Type: application/json` and
 
 ## Related
 
-- Browser Rendering billing path still uses the Worker’s internal client for
-  `page_to_markdown` — that is unchanged.
 - Developer docs (read-only) pattern:
   [cloudflare-developer-docs.md](./cloudflare-developer-docs.md).

docs/contributing/skill-patterns/cloudflare-developer-docs.md

@@ -1,22 +1,15 @@
 # Skill pattern: Cloudflare Developer Docs (`developers.cloudflare.com`)
 
-The **`cloudflare_api_docs`** builtin capability was removed in favor of **saved
-skills** (or one-off **`execute`** code) so Cloudflare-specific doc reading can
-evolve without shipping Worker changes.
-
 Use this pattern when you need API or product documentation from
 [developers.cloudflare.com](https://developers.cloudflare.com).
 
-## Prefer cheaper steps first
+## Preferred approach
 
-1. **Markdown-preferred `fetch`** against a path under the docs site (see
-   allowlist below). Matches what the old builtin did before any billed
-   fallback.
-2. If the response is still hard to use (heavy HTML, etc.), call
-   **`codemode.page_to_markdown`** with the same URL so Browser Rendering runs
-   only when needed (billed).
+Use **markdown-preferred `fetch`** against a path under the docs site (see
+allowlist below). Keep the skill focused on direct docs retrieval instead of
+adding site-specific fallback machinery.
 
-## Path allowlist (same policy as the former builtin)
+## Path allowlist
 
 Paths must:
 
@@ -77,26 +70,23 @@ Adjust `name`, `title`, `description`, and trust flags when saving.
 	const url = new URL(path, ORIGIN).toString()
 	const res = await fetch(url, { headers: { Accept: MARKDOWN_ACCEPT } })
 	const body = await res.text()
-	const contentType = res.headers.get('content-type')
-	const needsFallback =
-		contentType?.includes('text/html') && body.trimStart().startsWith('<')
-	if (needsFallback) {
-		return await codemode.page_to_markdown({ url })
-	}
 	return {
 		status: res.status,
-		contentType,
+		contentType: res.headers.get('content-type'),
 		markdownTokenEstimate: res.headers.get('x-markdown-tokens'),
 		body: body.slice(0, 500_000),
 	}
 }
 ```
 
+Callers should inspect `contentType` before treating `body` as Markdown. This
+helper returns the raw sliced response body plus `markdownTokenEstimate`; HTML
+responses are not auto-converted.
+
 For parameterized skills, use **`meta_save_skill`** **`parameters`** (e.g. a
 required `path` string) and read **`params.path`** instead of a hard-coded path.
 
 ## Related
 
-- Billed fallback: **`page_to_markdown`** (`packages/worker` coding domain).
 - Tracking: [issue #120](https://github.com/kentcdodds/kody/issues/120) (broader
   official patterns folder).

docs/contributing/skill-patterns/index.md

@@ -7,4 +7,4 @@ Worker-only primitives (billing bindings, secret storage, etc.).
 | Pattern                                                        | Summary                                                                                                         |
 | -------------------------------------------------------------- | --------------------------------------------------------------------------------------------------------------- |
 | [cloudflare-api-v4.md](./cloudflare-api-v4.md)                 | Call Cloudflare API v4 with secret-aware `fetch` to `api.cloudflare.com` (replaces removed `cloudflare_rest`).  |
-| [cloudflare-developer-docs.md](./cloudflare-developer-docs.md) | Fetch markdown from `developers.cloudflare.com` with an allowlisted path; optional `page_to_markdown` fallback. |
+| [cloudflare-developer-docs.md](./cloudflare-developer-docs.md) | Fetch markdown from `developers.cloudflare.com` with an allowlisted path and return a bounded preview. |

docs/use/search.md

@@ -31,7 +31,7 @@ To get **full markdown and schemas for one hit** (for example a capability’s
 
 Examples:
 
-- `page_to_markdown:capability`
+- `kody_official_guide:capability`
 - `user:preferred_org:value`
 - `github:connector`
 - `my-skill-name:skill`

packages/mock-servers/cloudflare/src/worker.ts

@@ -30,14 +30,6 @@ type DnsRecord = {
 	ttl: number
 }
 
-type BrowserRenderingMarkdownBody = {
-	url?: unknown
-	html?: unknown
-	userAgent?: unknown
-	rejectRequestPattern?: unknown
-	gotoOptions?: unknown
-}
-
 type DashboardEndpoint = {
 	method: string
 	path: string
@@ -98,12 +90,6 @@ const dashboardEndpoints: Array<DashboardEndpoint> = [
 		description: 'Create DNS record',
 		requiresAuth: true,
 	},
-	{
-		method: 'POST',
-		path: `/client/v4/accounts/${fixtureAccount.id}/browser-rendering/markdown`,
-		description: 'Convert a page or HTML snippet to markdown',
-		requiresAuth: true,
-	},
 	{
 		method: 'POST',
 		path: `/client/v4/accounts/${fixtureAccount.id}/email/sending/send`,
@@ -532,24 +518,6 @@ async function handleEmailSend(
 }
 
 async function routeApi(request: Request, env: MockCloudflareEnv, url: URL) {
-	if (request.method === 'GET' && url.pathname === '/__mocks/markdown') {
-		return new Response('# Mock markdown\n\nServed as markdown.\n', {
-			headers: {
-				'content-type': 'text/markdown; charset=utf-8',
-				'x-markdown-tokens': '8',
-			},
-		})
-	}
-	if (request.method === 'GET' && url.pathname === '/__mocks/markdown-error') {
-		return new Response('# Mock markdown error\n\nServer error page.\n', {
-			status: 500,
-			headers: {
-				'content-type': 'text/markdown; charset=utf-8',
-				'x-markdown-tokens': '9',
-			},
-		})
-	}
-
 	if (!isAuthorized(request, env, url)) {
 		return errorEnvelope(401, 10000, 'Authentication error')
 	}
@@ -596,43 +564,6 @@ async function routeApi(request: Request, env: MockCloudflareEnv, url: URL) {
 		)
 	}
 
-	const markdownMatch = url.pathname.match(
-		/^\/client\/v4\/accounts\/([^/]+)\/browser-rendering\/markdown\/?$/,
-	)
-	if (markdownMatch && request.method === 'POST') {
-		const accountId = markdownMatch[1]!
-		const payload = (await readJsonBody(
-			request,
-		)) as BrowserRenderingMarkdownBody | null
-		if (payload === null) {
-			return errorEnvelope(400, 1001, 'invalid JSON body')
-		}
-		if (accountId !== fixtureAccount.id) {
-			return errorEnvelope(404, 1002, 'account not found')
-		}
-		const hasUrl =
-			typeof payload.url === 'string' && payload.url.trim().length > 0
-		const hasHtml =
-			typeof payload.html === 'string' && payload.html.trim().length > 0
-		if (!hasUrl && !hasHtml) {
-			return errorEnvelope(400, 1003, 'Either url or html is required.')
-		}
-		const mode = hasHtml ? 'html' : 'url'
-		const sourceValue = hasHtml
-			? String(payload.html).trim()
-			: String(payload.url).trim()
-		const markdown = [
-			'# Mock Browser Rendering',
-			'',
-			`mode: ${mode}`,
-			`source: ${sourceValue}`,
-			...(typeof payload.userAgent === 'string' && payload.userAgent.length > 0
-				? [`userAgent: ${payload.userAgent}`]
-				: []),
-		].join('\n')
-		return envelope(markdown, { status: 200 })
-	}
-
 	const emailMatch = url.pathname.match(
 		/^\/client\/v4\/accounts\/([^/]+)\/email\/sending\/send\/?$/,
 	)

packages/worker/.env.example

@@ -29,10 +29,9 @@ CLOUDFLARE_EMAIL_FROM=reset@kody.dev
 # SENTRY_ORG=your-org-slug
 # SENTRY_PROJECT=your-project-slug
 
-# Cloudflare API / Browser Rendering / Email (optional; defaults to the public Cloudflare API, and `npm run dev` sets a mock URL + token for `page_to_markdown`, local email, and the internal API client)
+# Cloudflare API / Email (optional; defaults to the public Cloudflare API, and `npm run dev` sets a mock URL + token for local email and the internal API client)
 # Uses Authorization: Bearer <token> against /client/v4/* endpoints.
 # Reuse `CLOUDFLARE_API_TOKEN` for the live API. In mock mode `npm run dev` injects a local token.
-# page_to_markdown also needs CLOUDFLARE_ACCOUNT_ID when it falls back to Cloudflare Browser Rendering `/markdown`.
 # CLOUDFLARE_API_BASE_URL=https://api.cloudflare.com
 # Skip starting the Cloudflare mock during dev (use live API + token in `packages/worker/.env` instead):
 # SKIP_CLOUDFLARE_MOCK=1