1
Import the messy sources
Paste Reels, TikToks, YouTube videos, blogs, paper scans, or a whole saved-list export.
AI kitchen agents
The recipe layer your cooking agent needs before it plans dinner.
ChatGPT, Claude, and other assistants can invent meals. ReelsMeals gives them a better starting point: the recipes you actually saved, structured from videos, blogs, and paper into a cookbook an agent can read.
Card required for trial. Public agent discovery includes llms.txt, OpenAPI, and skill metadata; your trial unlocks the private cookbook you build from saved recipes.
Ask your assistant
Use my ReelsMeals cookbook. Find five weeknight dinners from recipes I saved from TikTok or Instagram. Prefer recipes under 35 minutes. Reuse ingredients where possible. Return: 1. the dinner plan 2. ingredients to buy 3. links back to the recipe cards 4. the original creator for each meal
Not another recipe box. A data layer for dinner.
Three steps turn scattered saved videos into something an assistant can actually reason over.
2
Structure the cookbook
ReelsMeals extracts ingredients, steps, timing, source links, tags, and creator context into searchable cards.
3
Let agents work from data
Assistants read ReelsMeals discovery metadata and public recipe APIs instead of hallucinating from vague prompts.
What this unlocks
The jobs that only work once your recipes are structured data instead of a folder of saved links.
Plan from recipes you already wanted
Start from your saved food videos, not generic recipe suggestions an assistant made up.
Prepare the shopping run
Structured ingredients are the layer an agent needs before it can merge lists and spot overlaps.
Cook from the source
The recipe card stays readable, and the original video is one tap away when technique matters.
Recipe managers compete on storage. ReelsMeals competes on the dinner outcome.
The market is not "where do I save a recipe?" anymore. It is "can my assistant help me turn the recipes I saved into food on the table?"