VegBatch

Cook once. Eat all week. Free vegetarian and vegan meal planning with an automatic grocery list — live at vegbatch.com.

Pick a week of recipes (or let it pick four that share ingredients), scale them, and get one aisle-sorted list in quantities you can actually buy: "3 cans chickpeas", not "1,275 g". Works offline in the shop. No account needed; an account adds sync across devices, a shared household list, saved shops and cook history.

What's here

data/            the dataset — 254 recipes, 508 canonical ingredients, alias table
pipeline/        everything that builds, tests and deploys the site (Node, no framework)
pipeline/site/   the web app source: one HTML shell, one app.js, shared lib/
site/            build output (gitignored) — `npm run build` in pipeline/

Start with pipeline/README.md. It covers the grocery list, the staples-vs-pantry distinction, the canonical ingredient layer, importing recipes, and the deploy chain. SCOPE.md is the original plan and data model.

cd pipeline
npm install
npm run build          # data -> ../site
node serve-site.mjs    # http://localhost:4174
npm test               # unit + import + site checks (serve-site must be running)

Adding a recipe costs nothing and calls no API:

node import-recipe.mjs https://some-blog.com/lentil-soup/ --add
node import-recipe.mjs recipe.txt --add

How the recipes were made

The collection began as 268 phone photos of one household's recipe cards and printouts, read into structured data by a model once, then reviewed by hand. Every recipe is vegetarian; meat broth was swapped for vegetable broth and the swap recorded. Ingredients and method steps are transcribed as written; descriptions are our own, never a source's prose. Where a card named its source, that credit is kept in source_attribution and shown on the page. The photographs themselves are not part of this repository.

Licence

• Code (everything under pipeline/, including the web app) is licensed under the GNU Affero General Public License v3.0. You may run, study, change and redistribute it, including commercially, provided your version — including one offered over a network — is released under the same terms with its source.
• Data (everything under data/) is licensed under Creative Commons Attribution-NonCommercial-ShareAlike 4.0. Use it, build on it, share it under the same terms, with credit, and not for commercial purposes.

Made with love by Tributary.