I’ve been making a version of this coffee cake for years, and it’s always done me right. Before I went paleo, it was made with a blend of gluten-free flours. For a while, I’ve made it with just almond flour and it’s come out great, but…
I’m a perfectionist when it comes to the things I eat. For example: I made and enjoyed plenty of paleo chocolate chip cookie recipes, but was not 100% satisfied until I created what I believe to be the perfect ones.
And I’m on to something: The perfect grain-free flour blend.
1 part coconut flour | 2 parts arrowroot starch | 4 parts almond flour
There is still more experimentation that needs to be done; I haven’t tried it in every recipe. But for this coffee cake, it’s golden.
Apple Cinnamon Coffee Cake with Walnut Crumb Topping
For the cake:
1/2 cup almond flour
1/4 cup arrowroot starch
2 tablespoons coconut flour
1/3 cup coconut palm sugar
1 tablespoon cinnamon
1 teaspoon baking soda
1 teaspoon baking powder
1/4 teaspoon salt
1 tablespoon butter, plus more for greasing the pan
2 eggs
1/2 cup sour cream or full fat Greek yogurt
1 teaspoon vanilla
1 cup grated apple (about 1 large apple)
For the topping:
1 1/2 cups walnuts (or pecans, if you prefer)
1/2 cup almond flour
4 tablespoons melted butter
2 tablespoons coconut palm sugar
1 tablespoon cinnamon
pinch salt
- Preheat your oven to 350° and grease a 8 x 8 baking dish.
- Make the topping: pulse the walnuts in a food processor 10-12 times or until they are course crumbs. Add the remaining ingredients and pulse 2-3 more times until combined. Set aside.
- Wipe out and dry the bowl of your food processor and add your dry cake ingredients. (almond flour through salt) Pulse a few times to mix.
- Cut the tablespoon of butter into smaller chunks and add it to the dry ingredients. Pulse 8-10 times or until it’s cut in to the dry ingredients, similar to if you were making a pie crust.
- In a small bowl, mix your wet cake ingredients (eggs through vanilla) and whisk until well combined. Stir in grated apple.
- Add to the food processor and mix until combined. Scrape down the sides once or twice to make sure it’s well mixed.
- Pour into the prepared baking dish and sprinkle the topping over, as evenly as you can.
- Bake for 30-35 minutes, or until a toothpick inserted into the center comes out clean.
- Allow to cool, and enjoy!








Is there anything I can successfully substitute for the eggs in this recipe?
I haven’t personally tried it, but have heard good things about using flax seed meal. 1 tablespoon ground flax seed meal : 3 tablespoons water, mixed into a slurry = 1 egg. Let me know how it works if you try it!
Thank you. I’ll let you know if it turns out well. Thank you for the recipe. It looks great!
I’m interested to know if the flax trick worked?
Doesn’t loading the ‘flour’ with starch make this more GF than paleo?
Does this recipe say paleo anywhere in it? It’s grain-free. Arrowroot is a starch, but is not a grain. It also has dairy, so if I were to label it anything other than grain-free, it would be primal, not paleo.
Thank you for sharing! Looks delicious, can’t wait to try it after I’m done with the sugar detox!
Simone, Thanks for sharing your ratio. I cannot wait to tinker with it in my grain free baking.
Question: Have you noticed that the size of the eggs you use make tweaking the coconut flour amount a necessity? Or perhaps you consistently use the same size eggs? [or measure them]
I find myself always needing to adjust my ‘flour’ amounts based on my egg size and I was wondering if I was the only one.
M
I find quick bread type recipes pretty forgiving as far as measurements are concerned. I get pastured eggs from the farmers market that vary greatly in size, (especially in the winter, when the hens aren’t all that excited about laying them.) so don’t really use them in baking. For recipes that I’m posting on here, I try to use the large eggs that I buy in the store, so that their size is more uniform. But it’s definitely a good thing to be able to recognize that your batter needs some tweaking!
Let me know how your grain-free baking comes out with the ratio I’ve been using! I’d love to hear…
I am definitely making this. I love your site. I am not a paleo, but I lean that way. Gluten-Free and learning as I go. I don’t do refined sugar and am working on cutting what I do use. I’ll be sharing if they are anywhere as good as they look. Thanks.
I wonder if you have any idea about the weights of the flours to use in your 1-2-4 flour formula. It seems like weighting the arrowroot and coconut flour, in particular, would help reduce variability and increase repeatability. If you don’t know the weights, how exactly to you measure?
You’ve got a good point there. I don’t weigh, but if I wanted to multiply any of my baked goods recipes, I would figure out the weights and scale by weight instead of volume. How do I measure? Honestly not all that carefully! I’m sort of a klutzy baker, and have found that quick bread type recipes are pretty forgiving. My skill lies more in the realm of savory creations, so I stick to the baking that’s in my comfort zone…
When you make a recipe some certain way you will measure the ingredients the same way each time. Even though you said “not all that carefully”, you aren’t sifting powdery ingredients into measuring cups one day and packing them in with a spoon the next. If you’re happy with your results then I don’t see why you’d need to change just because you wanted to scale the recipe. But where things get harder is when you’re passing your recipe to others. If you don’t give weights then other people may measure in a completely different way and the recipe may not work, or may give quite different results. I’ve heard that coconut flour can be a problematic ingredient to get wrong, and I expect it would exhibit the same variability as wheat flour: A “cup” of (wheat) flour can weigh anywhere from 4 oz to 6 oz depending on how you get it into the cup. So I was wondering: how do you get powdery ingredients into the measuring cup.
I scoop and level with a knife. I don’t pack. Does that help at all? When I have a better kitchen set-up (after the remodel I’m fantasizing about) I will try and measure as well, so people have a better gauge.
Genius! I’ve wanted an apple-based grain-free cake for a while now + now I won’t have to fumble around in the kitchen to figure one out. Baking tinkering is not my fave thing, so you are my hero!
Btw, so voting for your flour blend for president of the universe.
I’ll be following with interest the weight vs. volume saga as it unfolds.
Love it, Shelley, Thanks so much! And I’ll get myself a trusty kitchen scale soon
I didn’t know there was a saga.
I tried some things this weekend. Working entirely by weight I decided on a ratio of 5 parts almond flour, 2 parts arrowroot and 1 part coconut flour. (I like 8ths better than 7ths.) I also threw in xanthan gum—I don’t know if this makes a difference or not. I made a recipe for cut out cookies. In once case I replaced all the wheat flour with almond flour and in the other case I replaced it using the above mixture by weight. The texture and behavior of the dough with the mixture was much better than with straight almond flour, and the baked cookies came out better (though the pure almond flour substitution wasn’t bad). I made a chocolate chip cookie recipe using the substitution. In this case the cookies came out more cake like and less chewy than with wheat flour. I also experimented with Rose Levy Beranbaum’s Gateau Breton which is a kind of dense cake with an enormous amount of butter. I made a version with all almonds and a version with almonds and arrowroot. The version with arrowroot was 30% arrowroot and 70% almonds and had a distinct arrowroot flavor that my taster objected to. The plain almond version was pretty good. I made a third version using the mixture above, but haven’t cut into it yet so I can’t comment. ( Does the coconut flour magically cover up the arrowroot taste?)
It might… although I suspect that it does more for texture than flavor. I’ve made several recipes with around the same ratio (30/70 arrowroot/almond) and haven’t noticed a strong arrowroot taste. But every recipe is so different! You’ll have to let me know how the version with coconut came out.
Xanthan gum is absolutely only for texture. It’s supposed to be able to substitute for the gluten in wheat and provide structure to baked goods. It is often used in miniscule quantities. I have a pancake recipe I really like based on cottage cheese and whey protein where xanthan gum determines how thick the batter is, and hence how thick the cakes come out. If you leave it out you get really thin crepe like ones. If you put in 1/64 tsp you get a reasonable thin pancake thickness. At 1/32 tsp you get definitely thicker pancakes and my daughter complains they are too thick and she doesn’t want to eat them! As you get up past 1/32 tsp the batter gets too thick to spread out on its own when poured into the pan.
We ate the Gateau Breton cakes last night. I think that the 70/30 mixture cake had a definite arrowroot flavor that I didn’t like. In the 5-2-1 cake my taster claimed she still noticed the taste of arrowroot (but not coconut). The texture of this cake was better than the texture of the pure almond flour cake, and I didn’t notice the arrowroot flavor, which is interesting since it’s not that much of a decrease from the other one. The cake with only almond flour had a kind of oily texture. It’s possible this was because I undercooked the almond flour cake but not the 5-2-1 cake. But also maybe the arrowroot and coconut flour improved the texture by binding some of the butterfat. This cake is very plain, containing just butter, egg yolks, sugar and flour, so it’s probably more likely to expose the flavor of the “flour” than other recipes with more going on.
this is DELICIOUS! The texture is spot on! I would like to try it with Just Like Sugar or something as I am trying to avoid any sugar, but as it is its pretty perfect. Thank you!!!
You’re so welcome! I’m glad you love it! : D