Prototype on ideas.co.in · menufied.com and menuclature.com to be registered · every menu, price and number here is a sample unless the page says otherwise
menufied

Pitch deck · fourteen slides · print to PDF from the browser

Menufied, in fourteen slides.

Menufied · menuclature

The menu of record,
for every restaurant.

In the shape search engines and AI models read first. Built from the restaurant's own site. Sourced, dated, verified. Orders go to the official channel.

1 / 14

Canada's best-known brand cannot rank for its own breakfast menu.

  • "Tim Hortons breakfast menu": the #1 result is a near-exact-match site with no kitchen
  • It gets over 50,000 visits a month
  • Tim Hortons has zero pages targeting the query
  • Visitors try to order on the third-party site and cannot
50kmonthly visits to a copycat, one query, one brand
0pages on the brand site for the question
everychain, city and restaurant has the same gap
2 / 14

Why brands lose their own facts

  • Menus live in PDFs, images, JavaScript apps and ordering flows behind a location picker
  • There is no page per question: no "breakfast menu", "breakfast hours", "menu prices"
  • Answer engines cite what is structured, dated and attributable; brand sites are none of these
  • Third parties fill the vacuum with scraped, stale, unsourced copies, and still win
3 / 14

The vacuum is a market

  • Per chain: menu, prices, breakfast menu, hours, nutrition, allergens, vegan, kids, near me, delivery, coupons
  • Times every chain, every city, every local restaurant
  • Food first because the supply is the most public and the questions the most frequent
  • Then every industry: bank charges, telecom plans, hospital packages, school fees, gym prices
10+fact questions per brand
millionsof monthly fact queries in food alone, our estimate
1method that generalises: Menuclature
4 / 14

Product: three surfaces, one dataset

  • menufied.com: one page per chain per question, the menu of record, biggest button = order on the official channel
  • <restaurant>.menufied.com: a subsite per restaurant, generated from its own site and PDF, free, claimable, corrected by WhatsApp
  • Menuclature: the data layer: dish taxonomy, entity graph, prices over time, API, llms.txt, MCP, change feeds
5 / 14

Demo: the brand page

"Tim Hortons breakfast menu" on menufied.com: the menu board with dotted leaders and dated prices, breakfast hours, what changed since last month, the questions people ask, the official order button, and the same data as JSON-LD one click away. Open the sample page.

6 / 14

Demo: the subsite

anandbhavan.menufied.com, built from anandbhavan.in/menu.pdf on 11 September: the menu, hours, phone, map, the source capture and its hash, an extraction log, llms.txt and menu.json, and "claim this page". The restaurant's own site is a PDF nobody can read; this is the version search and models read. Open the sample subsite.

7 / 14

AI-first, by construction

  • schema.org Restaurant + Menu + MenuItem + Offer on every page
  • menu.json with source URL, capture hash, "as of", confidence and previous prices
  • menu.md and llms.txt for language-model crawlers; an MCP endpoint for agents
  • changes.json per restaurant; a city-level price index
  • Order path always official: site, app or phone
8 / 14

Why us

400Mdomain records with ownership: enumerate every restaurant site, tell chain from franchisee
yearsof captures: every menu has a history, every price a previous value
fleetrenders JavaScript and PDFs at scale; agents extract, normalise and check

Plus the names, the distribution of ideas.co.in and drops.co.in, brandnames.in for claims, and the Domain Fight Club to test it in public.

9 / 14

Business model, in order of arrival

  • Claimed-page Pro: what people ask about your menu, priority recrawl, the mark; small monthly fee
  • Data licences and API: answer engines, POS and delivery tooling, analysts; the price index
  • Sponsored placement only in the human view of lists; never inside a record
  • Never: orders, commissions, ads inside a menu, paid ranking of facts
10 / 14

Go-to-market

  • Phase 1: 200 largest chains in India, five question pages each, from their own sites
  • Phase 2: one city at a time through subsites; Bengaluru first, 2,000 restaurants
  • Phase 3: claims by WhatsApp; Pro for the claimed
  • Phase 4: API, MCP, licences; the menuflation index as the PR engine
  • Phase 5: the next vertical under Menuclature
11 / 14

Competition and the difference

  • Delivery platforms: menus for ordering, not the record; cannot publish without cannibalising
  • Google: shows what it finds; needs a source to cite
  • Menu aggregators and price sites: scraped once, undated, no source, no claim, dead end at ordering
  • Dated, sourced, verified, versioned
  • Official order path on every page
  • Subsite per restaurant, from its own site
  • Machine surfaces first: JSON, llms.txt, MCP, feeds

The full landscape, thirteen regions, seven kinds of player

12 / 14

Ninety days, three numbers

1,000chain question pages live and indexed by week 2
2,000Bengaluru subsites by week 6, claims counted per hundred
weeklysample of AI answers citing menufied by week 12; API keys issued
13 / 14

The ask

  • Register menufied.com and menuclature.com now, while both are free
  • Approve the 90-day plan
  • Choose the first city and confirm the first 200 chains from the corpus

Then the agents start on the chains the same week.

14 / 14