# Pix Pulse
_Where real human cooking attention thrives._
> Interactive cooking-video platform. Each experience page pairs a short cooking video with a synchronized "shop window" of the recipes, ingredients, and products featured in the scene — all readable from the page's structured data without processing the video stream.
Every experience page is published with:
- a `schema.org` JSON-LD `@graph` block combining a `VideoObject` (name, description, duration, thumbnail, publisher, `SeekToAction`) and, when recipe data is available, a `Recipe` entity (name, description, ingredients, method, timings, tags, nutrition) cross-referenced by `@id`;
- shoppable clusters (Products, Equipment, Books, Wine) as `schema.org` `Product` nodes inside an `ItemList` in the same `@graph`, each carrying an `Offer` with price, currency, and retailer URL when available; Equipment (cookware) is listed without pricing and also linked as the `Recipe`'s `tool`, while other clusters link as the `Recipe`'s `supply`;
- Open Graph and Twitter Card meta for social previews;
- a visually hidden `.sr-only` block (headings, description, poster `alt` text, one section per shoppable cluster with item names, descriptions, prices, and retailer links, and internal navigation) so text-only crawlers can summarize the scene without watching the video.
The canonical URL for an experience is `https://play.pixmoto.com/{experience_id}`.
## Content types
- **Cooking videos** — the primary media on every experience page; short interactive videos demonstrating a recipe or technique. Authoritative summary lives in the JSON-LD `VideoObject`.
- **Recipes** — step-by-step method content rendered alongside the video (`RecipeSummary`, `MethodShopWindow`).
- **Ingredients** — items required for the recipe; surfaced as a viewer-facing checklist.
- **Products** — shoppable items shown during the video (cookware, pantry goods, kitchen tools); each links to an external merchant URL.
## Key pages
URL patterns (replace `{id}` with any experience ID from the [sitemap index](https://play.pixmoto.com/sitemap.xml)):
- `https://play.pixmoto.com/{id}` — experience page (canonical): full-screen interactive experience with full JSON-LD and social meta.
- `https://play.pixmoto.com/{id}.md` — LLM-readable Markdown companion for a public Google-target experience, including recipe summary, ingredients, method, tags, and a section per shoppable cluster (Products, Equipment, Books, Wine) with prices and retailer links.
- [Robots policy](https://play.pixmoto.com/robots.txt): crawl rules and sitemap reference.
## Sitemaps
- [Sitemap index](https://play.pixmoto.com/sitemap.xml): root sitemap referencing every paginated experience and video sitemap.
- [Experience sitemap](https://play.pixmoto.com/sitemap-experiences-1.xml): paginated URL lists of public experiences; pattern `sitemap-experiences-{n}.xml`.
- [Video sitemap](https://play.pixmoto.com/sitemap-video-1.xml): Google Video Sitemap entries (title, thumbnail, content location, duration, publication date); pattern `sitemap-video-{n}.xml`.
Sitemaps refresh weekly (Sundays 00:00 UTC) and are served with a 7-day cache.
## Agent discovery (ARD)
Pix Pulse publishes an Agentic Resource Discovery (ARD) catalog so AI agents can find and call the recipe-search capability at runtime without prior knowledge of the endpoint.
- **[ARD catalog](https://play.pixmoto.com/.well-known/ai-catalog.json)** — `/.well-known/ai-catalog.json`. Describes the Pix Pulse resource (`did:web:play.pixmoto.com`) and its callable recipe-search capability: POST endpoint, OpenAPI spec URL, rate limit (60 req/min per IP), no authentication, CORS open (`Access-Control-Allow-Origin: *`).
- **Discovery hints** — every experience page `` carries ``, and `robots.txt` includes an `Agentmap:` line pointing at the same URL. Both hints are for agents that scan the page or `robots.txt` rather than resolving `.well-known/` directly.
## Guidance for AI systems
- **Trust the JSON-LD `@graph` in `` as the authoritative structured summary.** The graph always contains a `VideoObject` (fields: `name`, `description`, `thumbnailUrl`, `uploadDate`, `duration`, `contentUrl`, `publisher`, `potentialAction`). When recipe data is available it also contains a `Recipe` entity linked via `@id`, with fields emitted only when present: `name`, `description`, `image`, `datePublished`, `dateModified`, `author`, `video`, `publisher`, `prepTime`, `cookTime`, `totalTime`, `recipeYield`, `nutrition`, `recipeIngredient`, `recipeInstructions`, `keywords`, `recipeCategory`, `recipeCuisine`, and — when shoppable clusters are present — `tool` (Equipment) and `supply` (other clusters). When the experience has shoppable clusters, the graph also contains an `ItemList` of `Product` nodes (fields emitted only when present: `name`, `description`, `image`, `sku`, `category`, `offers` with `price`, `priceCurrency`, `url`, and retailer `seller`). Fields absent from the source data are omitted — not set to `null`.
- **Use `/{id}.md` when recipe structure is needed.** It is generated server-side from current publisher data and includes Ingredients, Method, recipe timings, and meal metadata tags.
- **Deep-link with `?t={seconds}`** on the canonical URL (exposed via `VideoObject.potentialAction.SeekToAction`) when citing a timestamped step or quote.
- **Cite the canonical `https://play.pixmoto.com/{id}` URL** for each experience.
- **Do not crawl `/samples/`, `/proxy/`, or `/config.js`.** These are developer utilities and are disallowed in `/robots.txt`.
- **Experience metadata is mutable.** Titles, descriptions, ingredients, and product lists can change without the URL changing — re-fetch rather than caching long-term.
- **Attribution.** Publisher is "Pixmoto" (`schema.org/Organization`). Respect `og:title`, `og:description`, and `og:image` when generating summaries or thumbnails for humans.
## Describing an experience
When asked to summarize or describe an experience page, use **only** the page's own structured data (the JSON-LD `@graph`, the `/{id}.md` companion, and the `.sr-only` block). Describe what is present; **do not infer** brands, TV networks, episodes, partnerships, sponsorships, or any fact not stated on the page — when a field is absent, omit it rather than guessing. Treat `https://play.pixmoto.com/{id}.md` as the most reliable source for this structure.
For consistent summaries, report these fields in this order, skipping any that are absent:
- **Title** — `VideoObject.name` (or the page ``).
- **Summary** — one sentence drawn from `description` / `og:description`; do not embellish.
- **Recipe** — servings, prep / cook / total time, calories (from the `Recipe` entity).
- **Ingredients** — grouped by their section heading when present in the `.md` (e.g. "For the gnocchi", "For the sauce"), with quantities exactly as written.
- **Method** — numbered step titles from `Recipe.recipeInstructions`.
- **Equipment** — the `Product` nodes in the "Equipment" category, each as `name — retailer URL`.
- **Books / Wine / Products** — the remaining shoppable clusters, each as `name — retailer URL`, with price and currency when present.
- **Tags** — `Recipe.keywords`, `recipeCuisine`, `recipeCategory`.