# Recipe Site pitch > A social cooking network built from the meals people actually make - HTML version: https://robbiepalmer.me/projects/recipe-site/deck - Project: Recipe Site (https://robbiepalmer.me/projects/recipe-site.md) - Slides: 12 ## Slide 1 # The family recipe book should learn from every meal Recipe Site turns cooking, adapting, and repeating recipes into useful evidence for the people you trust. ## Slide 2 ## The paper book loses the loop 01Findlinks, videos, family recipes → 02Cookscale, convert, time → 03Changeadapt to this kitchen → 04Repeatthe strongest signal Paper preserves the recipe. It loses what happened next. ## Slide 3 ## The useful personal product already exists Recipe boxCollection Search by ingredient… Chicken quesadillasMexican · 25 min Salt & chilli chipsBritish · side Capture & find URL, photo and Cooklang imports Cook modeStep 2 of 5 Fry the onions until soft. 2× portions08:42 Adapt & cook Scaling, unit conversion and concurrent timers KitchenStock match onionin stock tortillasin stock cheddaradd Plan & coordinate Meal plan, merged shopping and shared pantry Capture → plan → shop → stock → cook, in one working product. ## Slide 4 ## The market solved the checklist Tandoor Run the collection Migration, planning, shared shopping, offline use *(Interactive TechIcon component — view it on the HTML page)* Connect the food journey Communities, Made It, nutrition, retailer baskets *(Interactive TechIcon component — view it on the HTML page)* Win discovery Watch time, relevance, exploration Another long feature list is not a reason to build. ## Slide 5 The strongest signal happens after somebody taps save attentionwatched interestsaved intentplanned proofcooked fitadapted strongestrepeated Views measure attention. Meals measure whether a recipe worked. ## Slide 6 ## One score hides two different questions How good was it? → How well did it fit today? → Emergency fallbackuseful · not a favourite Tuesday staplegood · fast · reliable Skip next timeneither enjoyable nor useful Worth the work?great result · wrong occasion ## Slide 7 ## A recommendation should arrive with evidence EEllie recommended this"the kids actually ate this, zero complaints." weeknight · 25 minChicken quesadillasMade 128× · 54% made it again Save to boxAdd to planCook now Personal context, public proof, and an action that closes the loop. ## Slide 8 ## Four graphs answer "what should we cook tonight?" ```mermaid flowchart LR A[Recipe graph
versions and forks] --> E[Next meal] B[Cooking graph
cooks and repeats] --> E C[Trust graph
people who matter] --> E D[Kitchen twin
time, stock, diet] --> E E --> F[Cook · adapt · repeat] F --> B F --> A ``` ## Slide 9 ## The network grows from ordinary cooking 1Better fitA useful recipe reaches the right kitchen → 2More cookingThe household plans, shops, and cooks → 3Better evidenceRepeats and adaptations keep their lineage repeat No posting streak. No engagement chore. Dinner is the contribution. ## Slide 10 First finish the loop inside one household LiveCapture and cookRecipes, imports, scaling, timers LivePlan and stockMeal plan, shopping, pantry NextShare the stateServer-backed plans, lists, leftovers ThenClose the outcome loopVersioned cooks, repeats, adaptations ThenRank with evidenceTaste neighbours plus practical fit ## Slide 11 ## The build has an honest stop condition ### Use Tandoor If the real job is collection management, migration, planning, and self-hosted shared shopping. ### Keep building Only if cooks, lineage, trust, and kitchen context produce better next-meal decisions. If the network does not help real friends and families cook better, migrate the collection. ## Slide 12 # Build the network out of dinner, not engagement The next test is simple: do cooking outcomes and household context lead to better meals than ratings and feeds? [Try the product](/recipes) · [Read the full brief](/projects/recipe-site?tab=overview) --- Markdown index of this site: https://robbiepalmer.me/llms.txt