Show HN: I built a no-BS recipe search engine
stovetop.app
stovetop.app
Ever since I first saw this, whenever I get a recipe, from anywhere, I convert it by hand to this format.
Perhaps you can automate that process? That would be rad!
[1] http://www.cookingforengineers.com/recipe/108/Banana-Nut-Bre...
I recommend John Montroll's origami books.
1. Read and do all the "full-width" items at the top first as preparation
2. The remaining items in the tabulated recipe should be read as a combination dependency graph and Gantt chart. E.g. mashing ripe bananas can be done in parallel with melting the butter and beating the eggs, but all three must be completed in order to move onto the step of "mashing until smooth" all those ingredients together with the vanilla extract ingredient. This way of reading is more of a "left to right" approach.
https://modernistcuisine.com/recipes/silky-smooth-macaroni-and-cheese/Personally I find RxOL hard to read & edit though, so I have been thinking about whether it would be doable able to infer such tree structures from CookLang's more readable recipe format[1][2].
Ingredients and times are already marked, so surely verbs & follow up mentions of ingredients can be extracted from the text without further annotation? Haven't tried doing it yet though.
Just like with a Lisp, the benefit is that you always see at first glance what is the noun and what is the verb. With other recipes, you have to read through an entire “step“ instruction in order to parse that information.
Some of the comments on the site from ~2005 are so quaint ("Great site. Am linking to you!", etc.). Makes me feel like I'm in a quaint old-world town: proud of its anachronisms that morph into charm and wisdom with the passage of time.
Not exactly sure how the two compare.
Compared to Wilson's formula, it's very short, and it's not nearly as difficult to understand the idea behind how it works. One could easily get the question "Why would this random formula give better results than what a very established mathematician came up with?" I can't really answer that, but it would seem a lot of you agree that it does indeed produce better results when rating Steam's games. I can, however, try to give some insight into this.
For one, Wilson's formula isn't really meant to be used quite like this. It takes a rating and the sample size (the number of reviews), and outputs a confidence interval. And a confidence interval basically says that “We are some% sure that the score is between x and y”. If you increase the % of how sure you are, the distance between x and y also increases, and vice versa. But to get a single rating, it's not quite okay to just take the lower bound of that interval.
Secondly, because of what was mentioned in the last paragraph, it always gives us a lower rating than the original. This is clearly the incorrect behaviour, as something that just came out and gets a single negative review will be marked as having a score of 0%. Meanwhile, an established terrible game can have 10 positive and 500 negative reviews, and it will rank higher. This is also the reason why one of the two rules I listed was that all ratings should be biased towards the average.
Finally, while Wilson's formula probably gives us a more “precise” rating, so to say, it's not necessarily what we want to see. There's a lot of mathematics behind why what it does is correct, while the previously mentioned numbers of 2 and 10 that I picked for my formula were rather arbitrary. Still, I selected them so that the result would also account for the high number of reviews when assigning a good score. It's why you'll probably notice a lot less games with a low review count among the top games than before.
I think that's important because a game that is very popular and very highly rated should be ranked higher than a game that isn't as popular and is also very highly rated. Not because we can be more certain that this rating is indeed correct, but because you, as a random person who has yet to try that game, will more probably like it if a lot of other people have liked it as well — if it's not a niche game. And I think this aspect is definitely important and should be accounted for when trying to represent an entire game with just a single number.
A recipe that tastes bad should never be shown to anyone. But if there are two flavorful recipes and one of them has better instructions, then that's the one that should be sorted to the top. My go-to analogy for people being bad at documentation is to compare them to the variation in quality in cooking recipes. That's both a commentary on developer docs and on chef's docs.
I have been reworking a highly rated pie recipe. They have not covered browning the crust. They have made no mention of order and grouping of ingredients. If you follow the recipe literally, you're going to end up with a fluffy mix that won't fit in the pie tin. You're also going to get lumpy spices. That has a rustic appeal, but as the picture does not have lumps, the author is just bad at documentation. In my version, I split the spices so that you get a little texture but most are homogenized. I'm still experimenting with number of eggs. I've had 2 eggs (as in the recipe) taste eggy, and 3 eggs (minus a little egg white for the crust) taste fine, depending on how good the emulsion is, and that depends on order.
Most recipes still seem to be built on a model of fear and social barter. Fear that if Aunt Susie figures out Grandma Beth's brownie recipe, then nobody will 'have' to invite Grandma Beth to holidays or at least be excited to see her. If you want brownies then you need Grandma Beth.
Two days ago please, I want the cookies now.
I used to have a great recipe (I) called "15 minute cookies". The total time from first thinking about making them to eating - including preparation and cooking - was 15 minutes. Which was almost always irresistible, so I often made them. Somehow I lost the recipe. Does anyone know what that recipe might be? Thank you!
That one is pretty close. Just at step 4 instead of 30 mins she did 24 hours as the previous steps usually took 1-2 hours and she was tired because of how much flour she used. She also used 9-10 cups of flour instead of the 3-4 that calls for as she was also wanting to make a larger batch. If you did that little amount of flour it would probably be very and goopy to roll and cut in the next steps. She also fiddled the sugar, with corn syrup and molasses, I do not have it in front of me but I think it ended up being 1/2cup each and 1/2 cup regular sugar 1/2 cup brown. She also would add in 1/4tsp of anise. But basically all the wet stuff, spices into that, incorporate it good, then add flour until the correct consistency to roll out. With as much flour as she used you better have a good arm or a monster blender. But that recipe should be pretty close, there are thousands or variations on the internet (many include pepper and lemon), these fall on the ginger tasting side. They should be a crunchy cookie not doughy.
Is there anywhere that details the changes necessary to use for other rating scales (such as the 5-star/10-star system)?
1 stars = 0.00 upvote, 1.00 downvote
2 stars = 0.25 upvote, 0.75 downvote
3 stars = 0.50 upvote, 0.50 downvote
4 stars = 0.75 upvote, 0.25 downvote
5 stars = 1.00 upvote, 0.00 downvote
I hate this. It feels like part of the internet getting dumbed down.
If I search for "doughnuts", it's rather unlikely that I'm interested in "Chapssal doughnuts" (top spot), "Pumpkin Doughnuts" (second spot) or even "Gluten-Free Baked Chocolate Doughnuts Recipe" (third spot).
The list should be ordered (somehow, no idea how) by classic or common interpretation of the query first and pushing exotics, varieties and fusion stuff down to the very bottom.
Another one is "sourdough" - "Caesar Salad with Sourdough Croutons" is the top suggestion, followed by "Hard-Boiled Eggs and Parmesan on Toasted Sourdough" and "Radicchio Salad with Sourdough Dressing". Not exactly relevant.
It's a recipe app that can import from just about every recipe site/blog/etc and if it can't auto-import it then you can use the really easy tools to grab all the pertinent information. Once imported you can edit/scale/rate/tag/categorize/etc the recipe. Furthermore it has meal planning tools and a grocery list built in. I love the grocery list feature because you can easily click "Add to Grocery List" (of which you can have multiple if you want) and then just uncheck all the things you already have. It has a "Pantry" feature but I've never used it, I assume it will auto-uncheck items you already have when you go to add them to the grocery list but I'm not sure about that.
It's great to import a recipe and then tweak it after making it and see what works/doesn't so you don't have to find the recipe again later and/or remember the changes you made to it.
It's cross platform (Mac/iOS/Windows/Android) and running it on an older tablet in my kitchen is an awesome experience (timers built in, switch between multiple recipes easily, cross out ingredients after adding them, etc).
I have no connection to the company, I just love the app. If I had 1 request it would be a way to share recipes with friends through some "social"/"friend-ing" concept in the app (not using a social network). You can share a ".paprikarecipe" file that includes everything about the recipe but passing around a file isn't always easy and Discord just cuts off extensions longer than like 12 characters which makes it harder.
5 for ios. Free on android. And 30 for windows.
And seemingly no browser version? :(
As far as having to purchase per-platform, it’s a steal for the utility I get from it and there is no subscription fee, it’s 1-time.
We subscribe to a meal service called fresh prep which comes ad the problem from the other side, removing some prep and having to ask the question "What's for dinner". I think this is the killer feature Paprika is missing, browsing meal ideas that can be imported and added to your planner.
There is still lots of room for disruption in this space, thanks for sharing.
We tried several other apps and this one is worth every penny. Some of my favourite features: - Customize aisles - Auto stacking in shopping list - Clickable timers - Custom import from web - Well working scaling
Strongly recommending this app to anyone.
Sidenote: If available, use a portable scanner at the store. This reduces the amount of times you touch an item significantly, speeding up the experience.
Sidenote 2: To make auto stacking in the shopping list work the item name needs to be identical. We found that separating ingredient from processing works best. So 2 kg potatoes as ingredient and cutting in half in the directions instead of 2 kg of potateos cut in half as ingredient.
to me, no-bs means ingredients and instructions. Maybe a few pictures of the target meal. thats it.
That seems like pretty textbook (albeit very simplistic) mining to me.
The OP’s site is a really nice search engine, but it dumps you into the recipes’ unreadable SEO trash pit websites instead of parsing out the data
PlainOldRecipe isn’t working to strip recipes down for me
Making Coq-au-Vin, and use double the chicken? Probably not a problem. Add carrots, 0 carrots, totally fine.
Barbeque? You can do all sorts of things with the spice mix, it'll likely come out fine.
Making mac and cheese? 4x the cheese, totally fine, just keep adding it till it melts. Too thick? Add cream. Not tasty enough, add some bbq sauce. Measure? Why!
I commonly use a Pinch and a Dash of seasoning. Really it's tasting it and knowing what's off and being able to balance.
Because here's a real truth: every batch of a given ingredient likely has different strength. Dried, fresh, been in the jar for 2 years, brand new from the store, McCormick vs Store Brand vs Different part of the country/world... Also your vegetables will taste different depending on time of year, location, variety, as will your meats, dairy, cheeses, etc, literally everything.
Also, the fact that gasoline pumps are pressurised and standardise the volume based on a certain temperature should be a clue to one pitfall of volume-based measurements.
While its not quite right (its a little smaller), at the level of doing a recipe, 250ml [0] (but for flour in baking, you should use weight, not volume, anyway.)
[0] 4 cups = 2 pints = 1qt ≈ 1l
dear God please let this be satire, and make it satire if it isn't
A web search will turn up dozens of helpful sites that will do the conversion for you. Or the US could try joining the rest of the world.
“Why not simply…”
There's a LOT of variance in flour, milk and baking powder that make attempts at measurement based recipes a fool's errand.
The road to my hometown is called 363 so easy to remember. Just have to figure out which of them was the milk now again.
Metric and by weight is by far the best format for recipes and best way to measure ingredients.
In professional kitchens it's all we ever used (I used to work in some top restaurants, talking world's 50 best back in the day).
https://en.wikipedia.org/wiki/The_Boston_Cooking-School_Cook...
I've got cookbooks that date back to the 17th century. Lots of recipes from the 19th century. And old family recipes too. Doesn't mean I won't use metric and weights when I create new recipes, write recipes down, etc... And all my professional cookbooks are metric and by weight.
EDIT: Hey, wait, I actually found a good website with recipes: http://online-cookbook.com/goto/cook/rpage/000DDF
and a good search engine: https://search.marginalia.nu/search?query=pancake+recipe&pro...
if somehow you could combine the search of OPs site, plus the garbage stripping convenience of plainoldrecipe, plug in some really good unit conversion system, & save my preferences while you're at it, you might have the ultimate recipe tool.
This is the BS that I as hoping would be removed, but it was left intact unfortunately.
I've personally found mysaffronapp.com to be way better at this: It has been able to process recipes from every website I've shot at it. I think the developer contributes here to HN, as i learned about the site from a post she/he made.
https://addons.mozilla.org/en-US/firefox/addon/recipe-filter
Recipes are not copyrightable but if you pair a recipe with a story, the entire deliverable is copyrightable [0] because the recipe becomes a work of expression rather than a rote list.
> In other words, a bare recipe, without literary expression, is not copyrightable
[0]: https://www.lexology.com/library/detail.aspx?g=3741a5c0-f146....
Either way its total garbage.
You can try to diminish the real-world confusion and inconvenience of said business model by saying "cmon, it's not that bad, what's the big deal, why not let them make some money, why you gotta be such a stick in the mud?" That does nothing to reduce my or other peoples annoyance and lost time caused by said business model though. It's real, it does waste my time, and your time, and everyone else's time, it does inconvenience us, and it does make it harder to get the information we want.
Because of that, I'm glad to see these other websites arrive to replace blogspam recipe mills. I plan to use websites like http://www.cookingforengineers.com and https://stovetop.app/ search for most all my recipes going forward. Thanks OP for stovetop and thanks to other commenters here for posting more resources for us to use.
As i've built the app, I realized I almost never care about anything more than a rough estimate of the ingredients needed. Even then, I really only care about a few key ingredients. Here is an example (https://github.com/steven-p-walsh/menuplannercmd/blob/master...)
I primarily use it because it was the only app I could find with a good OCR mode for importing recipes from physical cook books.
To intrepid cooks I encourage getting a good cookbook, e.g. from Julia Child or Mark Bittman – where you can learn cooking fundamentals and techniques aimed to teach you how to cook self-sufficiently.
Break away from recipes and you’ll be dancing around your kitchen to your own culinary tune instead of recreating mediocre click bait.
No recipe in it has been bad. One or two have been "too much work to be worth it", but on average, the food has been good to excellent.
https://www.amazon.com/dp/B00X2E308K/ref=dp-kindle-redirect?...
A perfect balance of technical information & pragmatic recipes for the amateur cook to be very productive in the kitchen with minimal tools & ingredients.
Cook's Illustrated is great, and the food is always excellent, but it always seems to me that it's more work than you strictly need to do to get that quality.
Cook's Country recipes are about 90% as good for about 80% less work, in my opinion.
Cookbooks have editors, and the recipes are tested so they can be made with common ingredients. They want cooks to be successful!
Oddly I've found cookbook recipes to ALWAYS be better than online recipes. Book recipes tend to be shorter, clearer, and more successful. Online recipes are okay but sometimes don't come out the way I'd expect, they're more fiddly.
It's great to have options!
Your'e right, I've found some good content online. e.g. Chef John / Food Wishes has great recipes and videos. He also conveys some good tips & technique.
Example: too much salt in your soup? Add a little potato starch.
And yes, it's a much better way of doing things, plus helps you to think about the effect of each ingredient on the outcome.
Aside from those I don't really have go-to places, I tend to prefer to read through a few (i.e. just whatever top few results searching for the thing I'm thinking of making) for a general idea and inspiration of particular flavours, then 'wing it'. In future I'll try using Stovetop for that search :).
Edit: Oh yes and BBC Goodfood as sibling commenter pointed out. That's often where I search actually (because it's multiple author) rather than a general search engine, so if it were a Stovetop source I could definitely use it for that purpose.
I wanted to be able to search recipes from sources I knew I could trust. I also wanted a way to sort recipes by rating and include and exclude ingredients I already had.
Going forward, I aim to add more sources, better ingredient filtering (hierarchical) and more dimensions (cuisine, meal time, etc.)
Let me know what you think! One thing I struggled with was whether to includes mode with pictures or not. I can’t decide if the loss in info density is worth the benefit.
Building is the first step, sustaining is the second.
A picture is worth a thousand words. Figure out how to sustain the site and info density will become less of an issue.
(A few links showed 404 btw)
I would remove the author column personally, and maybe you could remove " reviews" in the reviews column (just show the number) and show the time as a number in minutes? If no time provided just show nothing.
If you add pictures you could have the option to show as a grid, similar to a file explorer. In the list view you could have a small picture with a bigger one when you hover?
I'm curious how you're parsing ingredients. I built a keto recipe search tool a few years ago[0], and I got so into the ingredient parsing problem that I spun that off into a separate service.[1] I still maintain the open source version, if you're interested in using that for stovetop.[2]
Almond and Dried Fruit Pilaf With Rotisserie Chicken
Freestyle Roasted Chicken Parm
etc. Still show up even with chicken, pork and 150+ other meat terms in the exclude list. But overall it's much improved.
Unfortunately can't find a way to bookmark the exclude list without a JS injection, and the source is obfuscated, so I'm probably not gonna bother figuring out how to do that.
... hm. I see that it does send out a promising looking API request that could be possibly turned into something bookmarkable. Looks like it has a handy array name too.
https://pastebin.com/nddrvGcy <= a start at meat filtering
I like the minimal design.
Please take this as constructive criticism, because I like your project overall, and want to only encourage you:
Please consider testing your project in older browsers, and including no-JS configurations in your testing suite.
It is easy, as a developer with a beefy machine with the latest and greatest installed, to overlook these possibilities, and it severely limits the accessibility of your site by older devices, slower connections, security-conscious users, and other edge cases.
Calendar: dinner with Anna and John next Friday. I know Anna is vegan and John hates hot spicy food. Go through a possible list of menus (not recipes) for the night, based on food that takes X time to be prepared. Get me a complete list of what I need for y amount of guests (and a cost estimation).
Add what I need to an online grocery store list making sure it arrives the day before at the latest (but not too soon since I want it fresh).
Another use case. I have a profile where I set I'm vegetarian, I don't like soups, blah and blah. I am hungry. I go to the fridge and I make a list of what I have and how much I have.
I now search for recipes with those ingredients that can be done in less than 40 mins.
Or, I'm preparing something that requires different timers: go and set them for me! Tell me that it's now time to put that thing in water if I want to prepare it tomorrow morning, or that I can now don't pay attention to this process during 45 minutes and alert me after that.
As I was saying, the recipe is really the easy part for me :)
2. 500 error when typing a quotation mark - needs escaping: https://api.stovetop.app/recipes?q=%27
{
"statusCode": 500,
"code": "42601",
"error": "Internal Server Error",
"message": "syntax error in tsquery: \"':*\""
}https://based.cooking/ might be another good source of recipes if you're not already using it.
Tagging by country / locale is a good idea. Another thing I'm currently working on is using a data structure of ingredients instead of a flat list. This structure will handle things like different names for ingredients and hierarchy (ie. no dairy will exclude milk and butter as well.)
The Guardian: https://www.theguardian.com/tone/recipes BBC Good Food: https://www.bbcgoodfood.com Jamie Oliver: https://www.jamieoliver.com Delicious Magazine: https://www.deliciousmagazine.co.uk/recipes/ Great British Chefs: https://www.greatbritishchefs.com
I should add that there are legal restrictions on scraping and re-use of content from third party websites; I haven't checked if the above permit this.
I don't like it, but it beats using something else for a volumetric measurement.
For recipes, stars often let me down as they seem to be more related to how the recipe worked and could be followed, not how good it tasted. And the flavors seem to be pretty bland.
That's not a tiny number but it's not large either.
"Ooh I can't wait to make this!"
Or:
"I LOVE making chicken like this!"
...and a five star review.
Toggle that slider up to 5k and it's just them sitting there at ridiculous numbers all pegged at 5 stars.
Edit: Also they should add a filter for "paywall" recipe sites. I haven't been able to find any of the top recipes as browseable.
1) Clean up the plural duplicates, e.g. peanut/peanuts, leek/leeks, carrot/carrots, etc.
2) I've never considered the author, is that common?
3) Mode to toggle pictures on would help scanning, cooking is very visual
4) Not simple to design, but some way to have either/or ingredients (e.g. peanuts OR cashews) could be useful
The thing I struggle with for pictures is where to put them, but I'll play around with some options. Maybe they can be behind a toggle.
I would be cool if you could search by ingredients or food allergies. Somewhere you can basically state the ingredients you have and then it'll suggest recipes, but I understand if you don't have all the recipes indexed by ingredients
Sorting the table too!
Also please add a facets like feature to avoid the user selecting combinations of ingredients that don't match any recipes.
Example: If you made recipe X, you should try recipe Y which uses 75% of those same ingredients, or recipe Z which has 80% of the ingredients in common. That way you can buy those core ingredients in bulk but still have something new and fresh for dinner.
The meal-prep dream for me is 10-15 recipes that taste good, are as distinct as possible, but have the majority of their ingredients in common.
To butcher a phrase, spice needs to be the variety of life in this scenario.
There are other bells and whistles I've added or plan to add, like factoring in diet restrictions or nutrition goals, but the core of it was to basically to find what I could cook without going to a store.
Or do we need a recipe search engine search engine?
Are their open source projects making progress instead of everyone independently reinventing the same wheel?
Are you able to give any insight into how this works behind the scenes, is it all manually input?
Bookmarked for future use.
Everything gets thrown into a Postgres database with a vanilla FTS implementation.
I can't find the name of it, sorry.
I fully support what you're doing. I am not fussed with someone's back stories about how their great grandmother's auntie's friend had her hair dyed when walking her dog to pick cherries... Blah blah... Just want the recipe and method
But unfortunately... Advert revenues are a thing.
"Unlock New York Times recipes and your personal recipe box with a free account."
These days I have all three of those recipe websites, and others, blocked in my SERPs block add-on in the browser.
I guess I unintentionally turned in to a recipe snob over the years. But I simply don't trust a lot of online recipes.
I would filter out the ones from Ready Steady Cook though
What we need is wikipedia for cooking.
Wikipedia has a 'first come first stays' policy for British/American spelling - applying that to recipes would be disastrous, giving me the final say on how to make the de facto custard, just because I created the page before anyone else?
But you can't achieve that with recipes, because they're not encyclopaedic entries, they're one person's opinion piece on how to make a nice <whatever>, like a blog post.
I suppose the slight flaw in my argument is that you can have competing encyclopaedia publishers - choose your namespace, your single source of truth, within which there's the one entry...
I searched for 2 recipes which are not particularly common - watermelon juice and scallion pancakes and there was an abundance of results, including ones I have used before.
We do not want your life story recipe person! We want to cook dinner!
small bug: go to page 2 of some search results and enter a different search term. You'll stay on page 2 and if there are not enough results to fill page 2, it will just be blank.
' breaks it.
Ok, I know this filter is not on ecommerce sites b/c money but one day there will be enough competition and then ...
Ok, I know there will never be real competition b/c capitalism. Why can't you just let me have this little fiction?
Ok, I know b/c reality.