So I started developing something remarkably similar about 4 years ago (in Django, too!). Wrote up a business plan with monetization, p&l, etc. I brought the plan and prototype to a few seed folks for early funding before I admittedly lost interest and moved onto something else.
The hardest technical challenge I had was, knowing there would need to be a ton of recipes in the system to make it truly effective, I had to do an automated intake of recipes from many places. But ingredient normalization got in the way, even with really good regex/etl practices. For instance, a recipe says, "boneless skinless chicken breast".. another says, "skinless boneless chicken breast". Some list the # of breasts. Some list pounds. Some mean the breast is split, some don't. But in order for the nutrition info to be accurate, the normalization process had to be near perfect. I ended up "buying" the source code to "recipefox" a recipe parsing plugin for firefox (for $100 or something like that) which helped tremendously, but still wasn't good enough for my tastes. I felt like to make it all work, i needed to essentially build a recipe ETL.
The monetization was kind of cool, I felt. Free for users. I felt that companies (food companies, supermarkets, etc) would put coupons for specific brands or specific stores on printed-out shopping lists based on the weekly recipes. i actually got alot of interest from that one from actual supermarkets. Thought that was cool. Good luck with this. It was a tremendous idea 4 years ago and I always wondered when someone else was going to do it because it just seemed so damn obvious to me.
"I have a gluten free child and a wife on a diet. What should I buy at the grocery store this week and what the hell should I make?"