How to Import Recipes from Allrecipes, NYT Cooking, and BBC Good Food

If you cook from the internet, you already know the problem. You bookmark a recipe on Allrecipes, save another to your NYT Cooking account, pin a third from BBC Good Food, and two weeks later you cannot find any of them. Each site has its own save button, its own app, its own folder structure, and none of them talk to each other. This guide walks through importing recipes from the three most trafficked English-language recipe sites into a single library, including the paywall and formatting gotchas that trip up most importers. ## Why these three sites are worth automating Allrecipes gets roughly 60 million visits per month. NYT Cooking has more than a million paid subscribers. BBC Good Food dominates UK search results for classic British cooking. Between them, they account for a huge share of the recipes home cooks actually use, and each one uses a slightly different underlying structure. ## Allrecipes: the easy case Allrecipes publishes recipes as standard schema.org Recipe JSON-LD. Every recipe page includes machine-readable ingredients, instructions, timings, yield, and nutrition data. Any reasonable importer can parse it cleanly. ### Step by step 1. Open the Allrecipes page you want to save. 2. Copy the full URL from the browser address bar, not a shortened social share link. 3. Paste it into the import box at /import. 4. Wait two to three seconds. The ingredients, instructions, photo, and total time should populate automatically. 5. Review the servings field. Allrecipes sometimes lists servings as a range ("8 to 10"). Pick the value you actually plan to cook. ### Gotchas Allrecipes shows user reviews inline with the recipe. A naive scraper will sometimes pull a highly rated review tip into the instructions. A schema-aware importer ignores the review block entirely, which is why the URL method is more reliable than copy-paste. ## NYT Cooking: handling the paywall NYT Cooking is subscriber-only for most recipes. If you are not logged in, the page returns a truncated version of the recipe with only the first few ingredients visible. Importers that fetch the public HTML get the stub, not the full recipe. ### Step by step 1. Log into nytimes.com in your browser first. 2. Open the recipe you want. 3. Use the "Print" option in the NYT Cooking page menu. This produces a clean, full-recipe view. 4. Either save the printable page as a PDF and upload it at /import, or copy the full text and paste it into the import box. 5. If you are a subscriber and want one-click imports, use the browser extension, which carries your authenticated session with it so the scraper sees the full page. ### Gotchas NYT Cooking uses em-dashes, fractional Unicode characters (½, ¼, ⅔), and writes temperatures in Fahrenheit only. A good importer converts the fractions to decimals and optionally flips the temperature to Celsius. If yours does not, check the ingredient list for "½" rendered as a broken box character and fix it before saving. ## BBC Good Food: metric, method-dense, generally clean BBC Good Food uses metric measurements by default, separates ingredients from "for the sauce" / "to serve" groups, and writes instructions as dense numbered paragraphs rather than short steps. ### Step by step 1. Copy the full URL (they look like bbcgoodfood.com/recipes/name-of-recipe). 2. Paste into /import. 3. Check the ingredient groups. If the recipe has subsections ("for the dough", "for the filling"), the import should preserve them as labeled groups. If it flattens them, edit manually. 4. Review instruction steps. BBC's "STEP 1" paragraphs often contain three or four actions. Split them into separate steps if you prefer a more granular read while cooking. ### Gotchas BBC sometimes writes "1 x 400g tin chopped tomatoes" where the "x" means "one tin of". A parser that treats "x" as multiplication will decide you need 400 tins. Verify tin/can ingredients after import. ## Bulk importing an entire recipe folder If you have been saving to all three sites for years, do not re-import one at a time. Export what you can: - Allrecipes: use the "My Recipe Box" export (CSV). - NYT Cooking: the site exposes a "Recipe Box" you can scrape via your logged-in session. - BBC Good Food: no native export; use the bookmark list with a bulk URL importer. Paste the resulting URL list into /import as a newline-separated batch. Most importers handle 20 to 50 URLs per batch cleanly; beyond that, rate limits on the source sites kick in. ## After the import The point of consolidating is not just having everything in one place. It is being able to search across all of it ("chicken thighs under 30 minutes"), scale by servings, and drop any of them into a meal plan without clicking across four tabs. Once your library has 200+ recipes, the search quality becomes the whole product. Start with /import for a single URL. If you have a backlog, batch the rest over a weekend. For exploration, /explore shows what's already in the public library so you can pull from there too. ## A note on schema drift All three sites have quietly changed their underlying HTML at least twice in the last three years. When an importer breaks on a specific site, it is almost always because the schema shifted and the parser has not been updated. If a recipe imports incomplete, the first troubleshooting step is to try again in 24 hours; most importers auto-patch their parsers within a day. If it still fails, fall back to the printable-page PDF method described in the NYT section. That workflow sidesteps scraping entirely. ## Quick reference | Site | Best method | Gotcha to watch | |---|---|---| | Allrecipes | Paste URL | Review text bleeding into steps | | NYT Cooking | Print to PDF then upload | Unicode fractions | | BBC Good Food | Paste URL | "1 x 400g tin" parsing | Start with the site you use most. If it is Allrecipes, a weekend batch job moves hundreds of saved recipes into a single searchable library by Sunday night. NYT Cooking takes longer because of the per-recipe print step, so amortize it over a month of evenings rather than trying to do the whole library at once.