I don't think I can offer much of an API. As I mentioned in another comment, I gave up on trying to parse a free-text recipe. I avoid a lot of headache by directing users to enter ingredients from the database list only.
That being said, I'm open to talk about it - see my HN profile for email.
edit: @phugold - I like the idea of your site, but you might want to take note of the ability to immediately see recipies without having to register. It's great to be able to just dive right in rather than hitting a wall requiring login.
After a lot of exploration I decided to limit the scope as much as possible and target home cooks and food bloggers, with a focus on recipes.
I spent a few weeks of my spare time trying to get the parsing of free-text recipes to work (essentially MFP's import from link). The results were decent, but I realized that if it's only 90% accurate I still need a lot of user interface functions (like MFP) to let users edit/fix the inaccurate bits. It started smelling bad so I decided to strip it down to its most basic function.
I have no plans for now to use databases of branded products - the ones I saw were quite expensive and would be more useful for sites like MyFitnessPal, perhaps less useful for recipes/cooking.
As a regular MyFitnessPal user I have to say, their recipe importer leaves a lot to be desired. Whenever I use it, I inevitably have to make corrections because it mapped the parsed ingredient to the wrong item. (And their UI for making corrections is not particularly great either.)
It's probably a good idea that you decided not to go down this road.
Yes I could sign up with an anonymous email address over a tor connection.
but I would resent that I was forced to do it
Someone even commented above about using a google/twitter/facebook integration. I'd be more willing to do it from there.
If you are a registered user of recipelab.org and have supplied your email address, recipelab.org may occasionally send you an email to tell you about new features, solicit your feedback, or just keep you up to date with what’s going on with recipelab.org and our products. If you send us a request (for example via a support email or via one of our feedback mechanisms), we reserve the right to publish it in order to help us clarify or respond to your request or to help us support other users.
Contrast that with:
When you register for recipelab.org, we collect certain personal information (ie - your email address and name). We will only use this email address to send you {monthly||weekly||daily} requests for feedback or updates on the product.
If you send us a support request, we may publish it to help us support other users. However, before we publish it, we will strip out any personally identifiable information.
That rubs me a little the wrong way too.
But a simple screenshot image would not work for all cases, because the calculator page renders differently on mobile devices. On the other hand, if I offer a full working version without sign-up, there would be no reason to sign up. If I show a "frozen" version of the full calculator, normal users might get the wrong idea from that as well.
It seems that you want to prevent people from experiencing your site without signing up first. Why is this? If you hide the value of your site behind the sign-up link, hardly anybody will bother signing up.
With a site like this, I would suggest enabling all features for anonymous users, except the ability to save recipes. Users can play around with the site and make their own recipe. The natural step after creating a test recipe would be to save it - this would be the best place to prompt for signup. I'd also recommend changing your on-boarding flow so that I do not need to confirm my email address immediately.
1) Search found all the ingredients I needed or pretty close approximations which was cool. Multiple units of measurement was an excellent choice. Food data looks USA - it would be helpful to know that as a UK user.
2) I had to make a username and password which will put off some potential users. Why not make it log in with Facebook/Google as an option?
3) I entered a quantity of tortillas and couldn't amend it without deleting the item. Was I doing something wrong?
4) As an active 6'2" man I suspect I need more calories than average. How about activity level/height/weight options? Kudos for having 'pregnant' and 'Breastfeeding' as options. However they would be more powerful as flags - my wife is both pregnant and breastfeeding! Or even 'add person' so one can look at nutritional info for cooking for a family.
I'm interested to see how this tool evolves! Good work!
1) Yes, it uses a USA gov' database. But I'm curious about why it looks American to you. Could you please elaborate?
2) To be frank, this is django's registration system, off-the-shelf (least effort). If I see enough interest in the site I will add simplified login. Thanks for highlighting it.
3) You did nothing wrong. I've stripped the interface down to the most basic functions.
4) You're totally correct. Adjusting the targets for body shape and activity is a priority item. As far as pregnant + breastfeeding, the source database has data for each but not both.
1) People in the enterprise space will be willing to deal with the weird nuances of the USDA-SR data. For example writing 'beef cooked' vs' 'beef raw'.
2) In the general consumer space, there is a huge problem in getting the general public to find ingredients they use. I found lots of people typing in brand name ingredients not matching what is in the USDA data set.
To me, the problem is partially solved with My Fitness Pal's huge dataset of user generated data. However the flip side is it is missing most of the micros as they are entered via a nutrition label.
Another solution might be to reference nutritionix.com 's dataset (I think the joyapp.com uses it also)
Unfortunately nutritionix has only basic nutritional data as well, the sort of thing you find on a product label. Not bad but not great. Have a look at their sample data. Plus it's around 10k/year or 50k/license.
I'm thinking home cooks would get a lot of value from the tool, and reasonably good data even if they use raw instead of cooked (as an example).
Try entering "potato fried" in the search box.
http://www.wolframalpha.com/input/?i=2+slices+of+bread+%2B+1...
