Salted Mocha Brownie Ice Cream
Dark chocolate and espresso ice cream with a fudgy brownie profile, using mild cannellini beans for body and fiber without intentionally tasting of beans.
Balanced on paper; not physically tested yet.
Ingredients
Tap ingredients as you weigh them.
Cannellini mocha-brownie base
The small print names the exact — often Dutch — products this recipe was calculated with. Equivalents work, but if a label differs materially, the formula deserves a recalculation.
Prepare
Rinse the beans thoroughly
Drain the canned cannellini beans, rinse them under cool running water until the water is clear and no foam remains, and drain very well. Blot away visible surface water, then weigh 85 g. This removes canning liquid and reduces the flavor most likely to reveal the beans.
Disperse the dry ingredients
In a completely dry bowl, thoroughly combine the sugar, dextrose, 0.2 g guar, and salt. Break up every lump and make sure the tiny guar dose is evenly diluted before it touches liquid. Do not add xanthan or another stabilizer.
Bloom the cocoa and melt the chocolate
Take 75 g from the measured whole milk and combine it with the 25 g espresso in a small saucepan. Heat until steaming but not boiling. Remove from the heat and gradually whisk in the cocoa until smooth. Add the finely chopped dark chocolate, leave it for 30 seconds, then stir until completely melted. Cool until no warmer than lukewarm.
Purée the beans completely smooth
Put the drained beans and remaining 100 g whole milk in a narrow blending jug or countertop blender. Blend for 60-90 seconds, scrape the vessel, then blend again until no skins, flecks, or graininess are visible. A countertop blender will usually produce the smoothest result. Do not freeze a visibly coarse bean mixture.
Finish the base
Add the cream, glycerin, vanilla, cooled chocolate mixture, and dry mixture to the bean purée. Blend for 45-60 seconds until completely homogeneous, keeping the blade submerged so the base does not become foamy. Scrape the sides and blend again briefly if any cocoa or bean paste remains.
Taste and age
Taste the cold base. It should read as dark chocolate brownie first, followed by restrained coffee and salt, and it should be slightly sweeter and more aromatic than the intended frozen dessert. There should be no obvious bean flavor or graininess. Cover and refrigerate for 4-12 hours, then stir or blend briefly if the base has thickened unevenly.
Fill and freeze
Pour the approximately 454.8 g base into the NC300EU pint. Confirm it remains below the molded MAX FILL line, fit the storage lid, and freeze on a level surface for at least 24 hours.
Spin plan
Confirm the frozen surface is level, there is no tall hard hump, the pint remains below MAX FILL, and the base is within -12°C to -25°C.
Run Lite Ice Cream once.
If creamy, stop. If safely powdery or crumbly, lower the outer bowl, scrape the sidewall and loose powder inward, make a small central well, add 5 g whole milk, and RE-SPIN once. If too soft, do not keep spinning; refreeze until firmer. Record whether the 5 g milk was needed.
Smooth leftovers flat before refreezing. When hard, rerun Lite Ice Cream rather than RE-SPIN; refreeze behavior has not yet been tested.
Notebook
Design brief
- Machine and container: Ninja CREAMi NC300EU, one standard 473 ml pint.
- Target: dense, fudgy dark-chocolate brownie ice cream with a restrained espresso note and enough salt to sharpen the chocolate; no intentional bean flavor.
- Serving plan: eat immediately after processing. Refreeze performance is not claimed.
- Assumed freezer: approximately -18°C, with a level freeze of at least 24 hours and a processable base temperature within -12°C to -25°C.
- Nutrition direction: meaningful whole-bean fiber and moderate protein, but not a low-calorie or low-sugar recipe.
- Available specialty ingredients used: dextrose, food-grade glycerin, and 0.2 g guar. Xanthan and stevia are intentionally excluded from the first test.
- Main test variables: complete bean-skin pulverization, absence of bean flavor, first-spin dryness, and whether the combined bean/cocoa structure produces an overly thick melt.
Development notes
JHermann’s Chico Caramel demonstrates the closest whole-bean precedent: its 680 g Deluxe formula contains 60 g chickpeas and 40 g aquafaba under cocoa, dates, caramel flavor, soy milk, soy protein, polyols, glycerin, and optional alcohol. JHermann reports good flavor but also describes the texture as dense, crumbly, and hard before revising the soy-protein and glycerin levels. It is therefore a research lead rather than a formula to scale directly.
This recipe is rebuilt for the NC300EU around AH Terra Cannellini bonen, whose current drained-bean label declares 6.8 g fiber and 6.9 g protein per 100 g. Cannellini beans are used instead of chickpeas because their mild flavor and creamy interior should disappear more readily behind cocoa and espresso. Rinsing, draining, and very fine blending are part of the flavor design, not optional cleanup steps.
The chocolate system uses Blooker cocoa and Delicata 70% dark chocolate, supported by AH whole milk and AH 35% cream. Sugar supplies familiar brownie sweetness, dextrose supplies additional freezing-point depression with less sweetness, and 5 g glycerin is a measured correction for the formula’s cocoa butter and unusually high solid load. No stevia is needed to hit the intended sweetness, and adding it would introduce a new aftertaste variable in the bean-concealment test.
Test notes
Not tested yet. Awaiting the first physical test batch.
- Cannellini brand, label, and drained weight method:
- Freezer temperature:
- Frozen-base temperature before processing:
- First-spin texture:
- 5 g milk and RE-SPIN required:
- Chocolate intensity:
- Espresso intensity:
- Salt balance:
- Bean flavor detectable cold or during melt:
- Bean-skin graininess:
- Fudgy body versus pastiness:
- Gumminess or overly thick melt:
- Sidewall ice:
- Immediate serving texture:
- Refreeze result:
- Digestive tolerance at the tested portion:
- Changes for the next batch:
Formula details PAC, POD, and solids for the curious
- Convention
- Sucrose = 1; includes added sucrose, dextrose, glycerin, estimated dairy lactose, dark-chocolate and bean sugars, and all label-equivalent salt
- PAC
- 28
- POD
- 15.4
- Total solids
- 40.3%
- Fat
- 8.4%
- Protein
- 4.4%
- Formula values describe the approximately 454.8 g frozen base before any optional 5 g post-spin milk. The 85 g drained cannellini beans are approximately 18.7% of the mix by weight.
- PAC uses sucrose FPDF 1.0/POD 1.0, dextrose monohydrate FPDF 1.9/POD 0.75, glycerin FPDF 3.7/POD 0.7, lactose FPDF 1.0/POD 0.2, and salt FPDF 5.9. Dark-chocolate sugar and the small bean-sugar amount are provisionally modeled as sucrose.
- The modeled PAC is approximately 28.0 including about 1.2 points from total label-equivalent salt, or approximately 26.8 with salt excluded. Cocoa butter and the high solid load can make the block behave harder than PAC alone predicts; the restrained 5 g glycerin dose is intended to counter that without dominating flavor or sweetness.
- The approximately 40.3% total-solids estimate counts drained-bean solids from the current AH Terra label, representative dairy solids, coffee soluble solids, cocoa, chocolate, sweeteners, glycerin, guar, and salt. Bean brand, drainage, and surface water can move the actual value materially.
- The 0.2 g guar is approximately 0.04% of the base. Cannellini starch and fiber, cocoa, chocolate, and dairy solids provide most of the body, so xanthan and additional stabilizers are intentionally excluded from the first test.
- Estimated fiber is approximately 16 g per pint, including about 5.8 g from the cannellini beans. Cocoa fiber is estimated because the current Blooker label does not declare it.
- Nutrition and composition are estimates based on current AH Terra cannellini, AH whole-milk, AH 35% cream, Blooker cocoa, and Delicata 70% chocolate labels plus representative dextrose, glycerin, espresso, guar, and vanilla values. Product substitutions require recalculation.
- Calculated and adapted formulation only: it has not yet been physically tested in the NC300EU.