Koji-Mold-Grown Block (KM-A) · GAIA R&D · single-step plant protein

§ 01 · WHAT IT ISOne ferment grows everything

Instead of separately making tempeh + tofu + rice and layering them, we mix a slurry from shelf-stable dry ingredients (rice flour, soy protein isolate, vital wheat gluten, hydrocolloids, dried umami stack), bind it with live rice koji + amazake, press it into a 3D-printed silicone mold, and let koji mycelium grow through the matrix for 6 days at 28-32°C × 78% humidity. The mycelium IS the binder, the saccharifier, the flavor developer, and the surface aesthetic — all in one step.

SHELF-STABLE DRY GOODS ───────────────────── rice flour · soy iso · gluten κ-carrageenan · TGase · shiitake kombu · salt │ ▼ ╔═══════════════════════════════╗ ║ MIX + WET WITH AMAZAKE ║ ║ + WATER → THICK SLURRY ║ ╚═══════════════════════════════╝ │ ▼ ╔═══════════════════════════════╗ ║ FOLD IN LIVE RICE KOJI ║ ← the inoculum ╚═══════════════════════════════╝ │ ▼ ┌───────────────────────────────┐ │ 3D-PRINTED SILICONE MOLD │ │ with 1.15mm grain channels │ │ ▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓ │ │ slurry packed firm │ └───────────────┬───────────────┘ │ CHEAP MURO 28-32°C 78% RH │ Day 1-2 ─ mycelium surface establishes Day 3-4 ─ penetration into channels + cross-link Day 5 ─ koji enzymes peak (Glu + sugars) Day 6 ─ HARVEST · demold │ ▼ ┌───────────────────────────────────┐ │ WHOLE-CUT BLOCK │ │ · firm sliceable │ │ · visible grain on cross-section │ │ · white mycelium "skin" │ │ · sears + steams + sous-vides │ └───────────────────────────────────┘

The breakthrough came from autoresearch Loops 1-8 (80 chemistry-scored variants, 11-Node framework). KM-A (composite 8.62/10) emerged from Loop 8 by:

Lowering humidity 85% → 78% — N5 surface-mycelium aroma jumped from 5.59 → 7.92
Raising gluten 39g → 49g — F18 disulfide adds structure for slicing
Pushing shiitake to 15g — N2 nucleotide ×30 synergy at ceiling
Grain channel 1.15mm — X6 grain visibility hits 10/10

§ 02 · VARIANTS3 top-3 paths from autoresearch

All 3 score above 8.5/10. Same family, fine-tuned along different axes. KM-A is the canonical first-trial. KM-B emphasizes gluten/structure (firmer slice). KM-C trades 0.1 composite for slightly lower cost.

KM-A · MASTER

8.62

L8-004 · cost RM 25.14
The chemistry-balanced converged variant. All 11 nodes ≥ 7.92. Use this for first kitchen trial.

N1 Free-Glu9.84 N2 Nucleotide8.30 N3 Aroma8.02 N4 Maillard7.92 N5 Surf aroma7.92 X1 Soft8.27 X2 β-glucan8.73 X3 Fiber8.80 X4 Water8.41 X5 Density8.64 X6 Grain10.00

KM-B · HIGH-GLUTEN

8.55

L8-007 · cost RM 24.80
Max gluten (50g) for firmest slice integrity. Best for thin-slice presentations (carpaccio, sashimi-style).

N1 Free-Glu9.48 N5 Surf aroma8.01 X3 Fiber8.80 X5 Density8.37 X6 Grain10.00

KM-C · COST-OPTIMIZED

8.51

L7-003 · cost RM 24.91
0.1 composite lower than KM-A but slightly cheaper at scale. Identical procedure. Useful for high-volume batch trial.

N1 Free-Glu9.41 N5 Surf aroma7.94 X3 Fiber8.80 X5 Density8.01 X6 Grain10.00

First-trial recommendation: KM-A If KM-A passes real sensory ≥8/10, then trial KM-B for slicing-test contrast. KM-C is the production-cost-optimized fallback once you've validated KM-A's profile.

§ 03 · CHEAP MURORM 280-535 — the kitchen R&D fermentation rig

A proper koji muro costs RM 25K-50K (controlledmold.com walk-in spec). For our R&D scale (1-3kg per batch), a styrofoam cooler with thermostat + heat mat + humidifier delivers identical conditions at ~1% the cost.

┌──────────────────────────────────────────┐ │ STYROFOAM COOLER 40-60L │ │ │ │ ┌────────────────────────────────────┐ │ │ │ HYGROMETER · Inkbird IBS-TH1 │ │ │ │ TEMP/HUMID display │ │ │ │ │ │ │ │ ┌─────────────────────────────┐ │ │ │ │ │ STAINLESS WIRE RACK │ │ │ │ │ │ (3D-printed molds × 4-8) │ │ │ │ │ └─────────────────────────────┘ │ │ │ │ │ │ │ │ Wet sponge tray (humidity) │ │ │ │ Reptile heat mat ───── outlet ──┼──┼──► Inkbird ITC-308 │ └────────────────────────────────────┘ │ (thermostat control) └──────────────────────────────────────────┘

Item	Spec	Cost (RM)	Source
Styrofoam cooler	40-60L, lid sealed	30-80	any hardware store / Shopee MY
Reptile heat mat	20×30cm, 10-15W, 28-32°C target	45-90	Shopee MY · reptile heat mat
Inkbird ITC-308 thermostat	plug-in temp controller, 0.1°C precision	90-140	Shopee MY · Inkbird ITC-308
Inkbird IBS-TH1 hygrometer	monitor temp + humidity, bluetooth log	60-95	Shopee MY · Inkbird hygrometer
Stainless wire rack	4" off floor, allows air circulation	25-50	Shopee MY · wire cooling rack
Mini USB humidifier (optional — wet sponge tray works)	50ml/hr output	30-80	Shopee MY · mini USB humidifier
TOTAL Phase 1 muro	capacity: 4-8 molds × 400g each	280-535

Critical setup steps before first batch

Sanitize interior with food-grade alcohol spray, let dry 30 min
Plug heat mat into Inkbird ITC-308 outlet — set target 30°C, dead-band 1°C
Place wet sponge tray (NOT touching heat mat) in bottom corner
Run empty for 12h with hygrometer to verify it holds 28-32°C × 75-80% RH steady
Only after stable: load molds onto the wire rack

§ 04 · 3D MOLD SPECThe IP-defensible alpha — grain-channel geometry

A 3D-printed PETG outer shell + cast-in-place food-grade silicone insert. The silicone has parallel grain channels 1.15mm deep, 0.8mm wide, 5mm apart. When koji mycelium grows through the slurry, it preferentially fills these channels → produces visible meat-grain on cross-section. Nobody in the koji-meat industry is doing this — MyForest grows in flat trays, Aqua Cultured uses extrusion.

PETG OUTER SHELL (rigid) │ ╔════════════════════════════════════════════════════════╗ ║ ║ ║ ╔══════════════════════════════════════════════════╗ ║ ║ ║ ║ ║ ║ ║ ┌─┐ ┌─┐ ┌─┐ ┌─┐ ┌─┐ ┌─┐ ┌─┐ ║ ║ ║ ║ │ │ │ │ │ │ │ │ │ │ │ │ │ │ ║ ║ ║ ║ │ │ │ │ │ │ │ │ │ │ │ │ │ │ 1.15mm ║ ║ ║ ║ │ │ │ │ │ │ │ │ │ │ │ │ │ │ channels║ ║ ║ ║ └─┘ └─┘ └─┘ └─┘ └─┘ └─┘ └─┘ ║ ║ ║ ║ ↑ ║ ║ ║ ║ 5mm spacing ║ ║ ║ ║ ║ ║ ║ ║ FOOD-GRADE SILICONE INSERT ║ ║ ║ ║ (FDA-21-CFR-177.2600 compliant) ║ ║ ║ ║ ║ ║ ║ ║ ◯ drainage hole ◯ ◯ drainage hole ◯ ║ ║ ║ ╚══════════════════════════════════════════════════╝ ║ ║ ║ ╚════════════════════════════════════════════════════════╝ │ Plate dimensions: 120mm × 80mm × 30mm (deep) = 400g final block Channels run PARALLEL to long axis → slice perpendicular to channels → reveals grain on cross-section

Item	Spec	Cost (RM)	Source
One-time capex
3D printer (entry)	Bambu A1 Mini or Ender 3 V3 SE	800-1200	Shopee MY · Bambu A1 Mini
Food-grade PETG filament	1.75mm, FDA certified, 1kg = ~30 molds worth	80-150	Shopee MY · PETG food grade
Food-grade liquid silicone	Smooth-On Mold Star 16 or Sylgard 184 medical · 500g kit	180-280	Shopee MY · food grade silicone Smooth-On
Per mold (reusable indefinitely)
Filament used per mold	~50g PETG outer shell	4-8	(amortized)
Silicone per mold	~80g cast insert	25-40	(amortized)
Print time	2-4h on entry printer	—	(amortized)
Cost per reusable mold	~RM 30-50, indefinite reuse	30-50

The 2-step mold making process (Formlabs / Prusa standard)

Print the PETG outer shell as a negative (the inverse of the final mold cavity) with grain channels embossed on the negative
Mix food-grade silicone parts A + B (1:1 ratio, platinum-cure), pour into PETG negative
Cure 4-8h at room temp → demold silicone insert
Sanitize silicone (boiling water 5 min, OR autoclave). Reusable 50-100+ cycles.
The PETG shell never touches food — only the silicone insert does

No 3D printer? Outsourced alternative JLCPCB / 3D Hubs / local makerspaces (Me.reka KL, MakerSpace USJ) print food-grade PETG at RM 30-60 per mold. Send them the .STL file (we'll author one in next phase) — they print + ship in 5-10 days.

§ 05 · INGREDIENT BOMKM-A canonical recipe · per 400g block

All ingredients are in the existing ingredient vault with verified Shopee MY supplier links. The big shift vs the V1/V5/V2 hericium master formula: no fresh tempeh, no tofu, no fresh mushroom substrate — only shelf-stable dry goods + live rice koji + homemade amazake.

Ingredient	Amount	Role	Source
Starch matrix + saccharification
Rice flour (fine)	123g	F2 starch matrix · bulk + Maillard reserve	any supermarket
Cooked rice (short-grain)	41g	F2 pre-gelatinized starch · koji food	cook fresh same day
Amazake (homemade)	77g	F4 ferment primer · instant Glu liberation	homemade — 6-10h recipe
Live binder + flavor developer
Rice koji (live, kome-koji)	91g	F4 LIVE INOCULUM · the binder + Glu/sugar liberator	Marukura Foods 500g
Protein body + cross-link
Soy protein isolate 90%	85g	F3/F19 protein body · isoelectric gel	Pharmanutri 1kg 91% halal
Vital wheat gluten	49g	F18 disulfide cross-link · structure for slicing	Halal Vital Wheat Gluten
κ-Carrageenan	5.9g	F2/F10 ion-bridge gel · X4 water lock	Halal Kappa Carrageenan
TGase (microbial Streptomyces)	0.5g	F11 cross-link · minimum dose (koji does most work)	WBS 500 TGase 50g
Umami stack (dry)
Dried shiitake powder	15g	N2 nucleotide × 30 amplifier · PMR7	Dried shiitake (grind yourself)
Kombu (powder or fine-cut)	4.8g	N1 free-Glu primary · PMR4	MOMMY Kombu Powder or grind whole 250g
Cure + hydration
Sea salt (non-iodized)	10.6g	~2.4% w/w · pathogen suppression + umami	any supermarket
Filtered water (60°C)	98ml	F-bound hydration · F12 disperse methylcellulose if added	boil + cool tap or bottled

Total dry mass: ~457g · final yield: ~400g block after mold drainage
Predicted COGS: RM 25.14 per 400g (≈ RM 12.57 per 200g portion)

Ferment conditions (the unlock from autoresearch Loop 8)

Temperature: 31.5°C (upper koji range — enzymes peak)
Humidity: 77.9% (the breakthrough — dropping from 85% to 78% boosted N5 surface aroma 5.59 → 7.92)
Duration: exactly 6 days (< 5 = under-developed, > 7 = bitter)
Mold grain channel: 1.15mm (X3 fiber 8.80, X6 grain 10.0)

§ 06 · 6-DAY PROCEDUREDay-by-day with verify checks

Day 0 Slurry build + mold load 45 min active

Step 1 (10 min): Sanitize 1L mixing bowl + wooden spatula + silicone mold (boiling water rinse + air dry). Pre-heat muro to 30°C + verify humidity at 78%.

Step 2 (5 min): Dry-mix into mixing bowl — rice flour 123g + soy isolate 85g + gluten 49g + κ-carrageenan 5.9g + TGase 0.5g + shiitake powder 15g + kombu 4.8g + salt 10.6g. Whisk to uniform.

Step 3 (5 min): Wet-mix in separate jar — amazake 77g + cooked rice 41g + filtered water 98ml at 60°C. Whisk to slurry.

Step 4 (10 min): Pour wet into dry, mix with spatula to thick uniform paste. Should pull a soft peak. No dry pockets.

Step 5 (5 min): FOLD in 91g live rice koji LAST — gentle folds, don't shear (preserves mycelium viability).

Step 6 (10 min): Press slurry into silicone mold, pressing FIRMLY to eliminate air pockets. Smooth surface. Brush an extra teaspoon of koji-water slurry over the top (surface inoculation).

Step 7: Wrap mold loosely in damp cheesecloth. Place on wire rack in muro. Verify temp 30°C + humidity 78%.

Slurry is uniform thick paste · no dry pockets · firmly packed · surface koji-brushed · muro stable at 30°C / 78%

If humidity reads >85% at start, open lid 2-3 min to vent. If <70%, add more wet sponges.

Day 1 Surface check + no opening 2 min

Read hygrometer + thermometer ONLY through lid window (don't open). Should hold 30 ± 1°C, 75-80% RH.

Through the cheesecloth/lid, surface should look mostly unchanged — koji still establishing.

Temp + humidity stable. Aroma faintly sweet-rice when sniffed near muro vents.

Day 2 First mycelium signs 3 min · brief peek

Briefly lift lid (10 sec max) — surface should show first wispy white mycelium fuzz. Aroma sweetening, slight koji-perfume.

If no mycelium visible by end of Day 2: temperature was too low. Bump muro to 31°C.

Visible thin white mycelium on surface. Aroma sweet-koji. No colored mold (green/black/pink).

If green or pink mold spots appear → abort. Sanitize entire muro before next batch.

Day 3 Mycelium thick · channel penetration begins 3 min

Surface fully covered in white koji "skin". Channels in mold filling visibly. Block firming up.

Aroma deepening — sweet-rice + earthy-mushroom + soy-savory all layered.

Full surface mycelium coverage. Block firm to touch through cheesecloth. Aroma multi-dimensional.

Day 4 Peak growth · enzyme cascade 3 min

Mycelium is at maximum density. Inside the block: koji amylase is converting rice starch → sugars. Protease is liberating Glu from soy proteins.

Block heat-output peaks today — may feel slightly warmer than ambient. Normal.

Surface mycelium dense + visible grain channels filled. Block holding shape firmly. Aroma at peak intensity.

Day 5 Enzymes saturating · flavor maturation 3 min

Mycelium activity slowing. Sugars + Glu + nucleotides peaking. Block firming further as proteins consolidate.

Plan harvest for tomorrow morning.

Aroma stabilizing — should smell like aged miso × fresh mushroom × sweet rice. NOT acidic, NOT alcoholic.

Day 6 HARVEST · demold 15 min

Step 1: Remove mold from muro. Unwrap cheesecloth carefully.

Step 2: Flex silicone insert sides gently → block should release. If sticky, run a thin spatula around edges.

Step 3: Place block on cutting board. Take photos (overhead · side · cross-section after first slice).

Step 4 (CRITICAL): Slice perpendicular to grain channels → reveal the visible mycelium-filled grain on cross-section.

Step 5: Cook test slice (sear, steam, sous-vide). Store rest in airtight container at 4°C — use within 3 days, or vacuum-pack + freeze for 60 days.

Block holds shape. Cross-section shows distinct white mycelium grain pattern. No off-smell. Sears with clean Maillard browning. Texture: firm but yields to bite.

§ 07 · SENSORY SCORINGValidate vs autoresearch prediction

Score each node 1-10 after Day-6 harvest + cook test. Compare against KM-A's predicted scores. Composite should land within ±1.0 of 8.62 — wider divergence means a recipe variable drifted (most often: humidity or grain channel depth).

11-Node Sensory Card · KM-A trial

Node	What to assess	Predicted	Your score
N1 Free-Glu	Umami body on first chew · savouriness	9.84	___ /10
N2 Nucleotide	Mushroom-stock depth · M1 ×30 synergy	8.30	___ /10
N3 Aroma volatiles	Earthy mushroom + ferment perfume on slice	8.02	___ /10
N4 Maillard	Sear crust browning · toasted-bread notes	7.92	___ /10
N5 Surface aroma	Koji-skin perfume on whole block	7.92	___ /10
X1 Softness	No rubber, no cloth-mouth · yields to bite	8.27	___ /10
X2 β-glucan x-link	Cohesive bite · no falling apart	8.73	___ /10
X3 Fiber anisotropy	Visible parallel fiber on tear · "muscle-grain"	8.80	___ /10
X4 Water gradient	Juicy interior · drier crust	8.41	___ /10
X5 Network density	Springback after compression · slice integrity	8.64	___ /10
X6 Grain visibility	VISIBLE white mycelium grain on cross-section	10.00	___ /10
COMPOSITE (avg of 11)			predicted 8.62 · yours ___ /10

Photo capture every trial

Overhead whole block (post-demold, before slicing)
Side profile showing mold grain channels filled with mycelium
Cross-section after FIRST slice — the visible mycelium-grain on cut face
Sear test plate — golden-brown crust on slice
Bite test — pulling slice apart, photo the strands + grain direction

§ 08 · TROUBLESHOOTINGFailure modes + quick-fixes

Green / black / pink mold on surface by Day 2-3 Contamination — likely from un-sanitized equipment OR humidity creep above 90% (condensation droplets land on surface, wild mold takes hold). ABORT the batch. Next time: sanitize all surfaces with food-grade alcohol, verify hygrometer accuracy, ensure cheesecloth is clean.

Block smells alcoholic or vinegary by Day 4 Temperature dropped below 26°C at some point → wild yeast / Lactobacillus took over from koji. ABORT. Next time: verify Inkbird ITC-308 holds 30°C steady, do not place muro near windows/AC vents.

No mycelium visible by Day 2 Either: (a) rice koji was old / inactive — verify spore viability with a small dish test, OR (b) temperature too low — bump muro to 31°C, OR (c) humidity too low (< 65%) — surface dried before mycelium could establish. Recover by bumping temp + humidity. May still succeed by Day 3.

Surface mycelium yes, but grain channels empty (Day 4-5) Channel depth too deep (> 1.5mm) for mycelium to penetrate. Next mold print: reduce channel depth to 1.0-1.15mm. Current batch may still be usable if X6 visibility scores 6+ at Day 6.

Block crumbles when slicing on Day 6 X5 network density too low — likely gluten was below 40g or TGase below 0.5g. Increase gluten to 50g for next batch (KM-B variant). May also indicate ferment went too long — try 5.5 days instead of 6.

Block holds shape but rubbery on bite X1 softness too low — TGase over-cross-linked. Reduce TGase to 0.3g next batch. Or κ-carrageenan was too high (> 7g) — reduce.

Bitter aftertaste on Day 6 Over-fermented (likely 7+ days OR temp too high > 33°C). Harvest at exactly Day 6, verify muro temp doesn't drift above 32°C.

§ 09 · USE + SERVEHow to present the demolded block

Cooking methods (best to worst for showcasing the grain)

Slice + sear

180°C cast iron · 60-90 sec per side · brings out N4 Maillard, preserves visible grain. Use a thin slice (8-10mm) for max contrast.

Sous-vide reheat

62°C × 20 min · then sear briefly. Best for preserving X5 network density (no drying out). Premium plating method.

Steamed

85°C × 8 min · for soft preparations (bao filling, congee topping). Loses some N4 Maillard but keeps tender bite.

Carpaccio (raw cold)

Thin-slice 2-3mm · serve cold with citrus + olive oil. Shows X6 grain visibility most dramatically. Note: only safe if Day 6 surface mycelium swab passed pathogen test.

Confit / oil-poached

85°C oil bath × 30 min. Adds richness, preserves shape. Best for "vegan duck confit" plating.

Cross-brand applications

Pinxin Rendang Hericium Mushroom — substitute the KM block sliced + braised in rendang sauce. Speeds the Pinxin label-extension play.
MIRRA premium bento — sliced sous-vide KM block as the "centerpiece protein" alongside grain + greens. Position at RM 38-48 premium tier.
Wholey Wonder tasting menu — 3-cut presentation showing the cross-section grain as the visual story.
Drstan fitness positioning — high-protein (22-26% w/w) lower-cost SKU vs whey or pea-protein products.

Storage + shelf life

Refrigerated 4°C

3-5 days in airtight container. Best within 3 days for peak texture.

Vacuum-packed 4°C

10-14 days with food-grade vacuum sealer.

Frozen -18°C

60-90 days. Thaw in fridge 24h before use. Slight X1 softness loss after thaw (drops ~0.3 on sensory).

Vacuum + blast-frozen -40°C (Pinxin infra)

6-9 months. Pinxin's existing blast-freeze setup is ideal for this product.

→ Back to R&D Workbench Ingredient Vault → Other ferment recipes →