Indonesia · Operational · 8,493 villages indexed · Genesis pipeline live
Energy Intelligence · Rural Electrification

From a coordinate. To a built village. Automated.

Indonesia's #1 rural electrification intelligence platform. The work that took twelve months and one hundred thousand dollars per site — feasibility, engineering, procurement, grant matching, construction operations — is now a two-minute API call. Across the entire archipelago. On demand.

8,493 unelectrified villages indexed — every kabupaten, every provinsi.
LIVEEngine pass complete · Desa Beleme · Papua Pegunungan · 47 buildings · 124 kWp · 4.35 PSH · IRR 18.4% · 1m 42s
Three readers, one platform
For Developers → For Government → For Capital →
Run the Engine on a Village See How It Works
BAPPENAS· PLN· ESDM· Kementerian Desa PDT· PT SMI· BPDLH· BPS· KEMENKEU· World Bank· ADB· IFC· KfW· JICA· MIGA· FMO· GCF· USAID· Verra· Gold Standard· ICVCM· BAPPENAS· PLN· ESDM· Kementerian Desa PDT· PT SMI· BPDLH· World Bank· ADB· IFC· KfW· MIGA· GCF·
0
Villages indexed, Indonesia
38 provinsi · 514 kabupaten
<2 min
Coordinate to bankable PDF
Median, production hardware
0
IEC/PUIL validation rules
Per engineering package, signed
0
Named funding instruments
$330M+ visible TAM
±1.5%
PV simulation accuracy vs pvlib
Better than PVsyst, browser-native
The Bottleneck

Feasibility studies are the bottleneck.
Not capital. Not generation.

A bankable feasibility study costs USD 80,000–150,000 per site, takes 12–18 months, and consumes six to ten specialists. Across 8,493 villages, that is USD 850 million and fifteen years of analyst time. No ministry has the budget. No developer has the runway. No financier will fund the studies. This closed loop has kept rural electrification under-deployed for thirty years.

01

Data Fragmentation

Building footprints, solar resource, household ability-to-pay, political risk, climate hazard, grant eligibility — spread across 30+ disconnected sources, four ministries, three multilaterals, two satellite vendors.

02

Per-Site Fixed Cost

A 500 kWh minigrid carries the same overhead as a 50 MW plant. The arithmetic kills aggregation before it begins. Distributed energy will not finance itself site by site.

03

Format Asymmetry

One project, six required formats — IFC, ADB, KfW, JICA, MIGA, Verra. Existing tools produce one document. The other five are written from scratch, by hand, every time.

04

The Aggregation Paradox

Financiers require minimum tickets of USD 10–25 million. Distributed sites are USD 0.4–1 million. Without aggregation, capital cannot deploy. Without per-site preparation, aggregation cannot happen.

What the alternatives can — and cannot — do.

None collapse the marginal cost of producing a bankable artifact. None solve format asymmetry. None operate at portfolio scale. None ship a sealed engineering package with HMAC audit chain.

CapabilityGIS DashboardsPVsyst / HOMERConsultanciesDonor PPFsLumen8 Engine
Spatial visualization✓partial✓✓✓
Techno-economic model—✓✓✓✓
Per-HH electrical design——partial—✓
30-rule IEC/PUIL validation————✓
Construction operations stack————✓
Political & ESG & grant logic——✓✓✓
Multi-format DFI output——partial✓✓
HMAC-signed audit chain————✓
Per-site cost$0$5–50K$80–150K$30–100Knear-zero
Cycle timeminutesweeks12–18 months12–24 months<2 minutes
Portfolio aggregation———partial✓
Digital MRV / Article 6 ready————✓
Multi-country adapter—partialpartial—✓
What Lumen8 Is

One engine. One dataset.
One evidence trail.

That triangle is the product. The engine compresses 18 months of multi-disciplinary feasibility work into under two minutes. The dataset is the only one of its kind that covers all 8,493 unelectrified villages in Indonesia at building resolution. The evidence trail is what lets a financier compress 12 months of due diligence into 4 weeks.

Layer 01 · The Engine
Genesis → Engineering → Operations

A 7-stage Genesis pipeline produces a faithful 3D digital twin from a single coordinate. A per-household engineering kernel converts the twin into a sealed construction package with 30 IEC/PUIL validation rules, dual-mode SLDs, TCC selectivity, arc-flash analysis, and PE-approval workflow. A Stage-4 operations layer carries the chain from APPROVED → ISSUED → AS-BUILT.

7 stagesCV terrainLV Steiner tree3D twin30 IEC rules12 ops skills
Layer 02 · The Dataset
Indonesia, indexed at building resolution

50.9 GB of curated baselines. All 8,493 unelectrified villages joined to Google Open Buildings v3 footprints, 168 months of nightlight history, the full RUPTL Lisdes 2024–2033, every ESDM substation, BNPB hazards, ACLED conflict events, BRWA adat territories, household ability-to-pay calibration. Why no one else can do it at this resolution.

8,493 villages50.9 GBGOB v3168mo nightlightRUPTL 2033CMIP6BRWA 2,208 adat
Layer 03 · The Evidence
HMAC audit chain & bidirectional source trace

Every metric — bankability grade, IRR, climate AAL, grid arrival year, political risk score — ships with a machine-readable evidence chain signed with HMAC-SHA256. Every contributing source, timestamp, confidence interval, and disagreement flag. A reviewer audits the chain, not the conclusion. Compresses 12-month due diligence to 4 weeks.

Merkle treeHMAC-SHA256PE-approval6-state machineBidirectional trace
70–80%
Projected close rate vs 35% industry baseline
4 wks
E&S review from 6–12 months
2.3%
PRI pricing vs 5–8% generic band
~$0
Marginal study cost (from USD 100K)
The Engine, Step by Step

From a single coordinate.
To a bankable digital twin. In under two minutes.

A seven-stage autonomous workflow that converts a four-tuple village identifier into a complete, spatially verified, financially modeled energy system. Every stage is auditable. Every output is anchored to upstream evidence. Scroll to watch the engine run.

01
AOI Frozen
"Lock the area of interest. Gate the run."

Spatial lookup against the 8,493-village index. Bounding box drawn at 800m radius. Three gates check before fetch starts: minimum 30 buildings, 80% WKT polygon quality, 800m max scatter. Auto-retry at 2× then 4× bbox on failure. Halts with actionable error if rural sparseness violates gates.

province / regency / district / village800m bbox3 quality gates
Source · js/genesis/engine.js · ADR-001
02
Data Fetch (parallel)
"Six layers, one bbox, in parallel."

Six external sources fetched simultaneously: Sentinel-2 L2A imagery (planetary computer, <20% cloud), Esri World Imagery (z=18), Google Open Buildings v3 footprints, OSM road graph (Overpass), Copernicus DEM 30m, BNPB hazard layer. In-flight deduplication, IndexedDB cache, 5-min negative-cache backoff after 3 failures.

6 parallel fetches13-adapter API fabricIndexedDB cache
Source · js/data/api-fabric.js · js/data/adapters/*
03
CV Verify
"100% real footprints, no synthetic fallback."

GOB polygons cluster-checked against threshold: median pairwise distance ≤ 800m, ≥ 30 buildings, ≥ 80% WKT well-formed. Outliers flagged. The strict gate enforces ADR-039: never fall back to synthetic geometry. If GOB lacks coverage, the village halts with a precise error and analyst follow-up.

GOB-first policyno synth fallbackADR-039
Source · scripts/bake-genesis/server.py · ADR-002 · ADR-039
04
Layout Canonical
"Per-building archetype. Per-HH peak kVA."

Each polygon classified by shape (rectangular / L / T / V), region, and OSM building:religion tag into archetype: residence, mosque, school, medical, market, government, other. Archetype determines peak kVA and daily kWh from the load model: kva_peak × 0.9 × 4 hours/day. Round-robin A/B/C phase assignment minimises imbalance.

7 archetypespeak kVA per HHA/B/C phase balance
Source · js/genesis/building-detect.js · render/archetypes.js
05
LV Design
"Steiner tree + voltage-drop convergence."

Steiner-tree solver finds optimal pole placement connecting every household. Cable CSA pre-sized per IEC 60364-5-52 three-criterion: ampacity (§523), voltage drop (§525), thermal stress (§434.3). Three-iteration convergence loop refines CSA. 3D drop lengths inflated using DEM elevation — +5–15% on Papua highlands. Cumulative ΔU reported per feeder, capped at 5% per PUIL 2020 §3.16.

Steiner treeIEC 60364-5-523-iter convergencePUIL 2020
Source · js/genesis/lv-designer.js · scripts/engineering/kernel/cable_solver.py
06
BOQ + 30-Rule Validation
"30 rules, zero errors, signed."

Bill of Quantities resolved against a 170-SKU catalogue with 9-region Indonesian pricing (Java vs Papua Pegunungan). 30 named validation rules execute (DESA-V-001..034) covering cable CSA, OCPD ratings (Icu ≥ Icc), bus voltage, RCD final circuits, PV string Voc cold ≤ 1000V, PE conductor on every cable, twin provenance ≥ 50%. Pass/fail badge. PE-approval gate if errors.

30 rules170-SKU BOQ9-region pricingHMAC-signed
Source · scripts/engineering/kernel/validation.py · ADR-038 · ADR-039
07
3D Digital Twin
"Faithful to the satellite. Built to be built."

Three.js scene assembles: terrain mesh from DEM, satellite-imagery-draped ground, extruded buildings on real WKT footprints (no clustering, no synthesis), PV array placed by science-based siting, LV poles + drop cables, hazard overlay. Procedural archetype meshes with LOD variants. Horizon shading baked (72-bin per-village) for downstream PV simulation. Interactive: orbit, layer toggles, zoom-to-building.

Three.jsWKT-faithful72-bin horizonLOD variants
Source · js/genesis/scene3d.js (99 KB) · ADR-009 · ADR-010 · ADR-013
Stage 01 · AOI Frozen
✓ 30 buildings minimum ✓ 80% WKT polygon quality ✓ 800m max scatter lat -4.082 lon 138.624 Desa Beleme Papua Pegunungan
PARALLEL FETCH · 6 LAYERS Sentinel-2 L2A · planetary computer 412ms Esri World Imagery · z=18 186ms Google Open Buildings v3 · 47 polygons 1.18s OSM Overpass · road graph 740ms Copernicus DEM 30m · elevation 312ms BNPB InaRISK · multi-hazard layer 528ms TOTAL ELAPSED 1.18 s Bottleneck: GOB v3 polygon fetch · in-flight dedup active api-fabric.js · 13 adapters · IndexedDB persistence · 5min neg-cache
GOB v3 POLYGONS · 47 RETURNED · CLUSTER CHECK outlier median pairwise = 408m ✓ CLUSTER CHECK PASSED 47 buildings · 92% WKT · 408m median scatter · 1 outlier flagged for review
PER-BUILDING ARCHETYPE · PEAK kVA · PHASE A/B/C A B C A B C A B C A MOSQUE SCHOOL B C MEDICAL A B MARKET C residence (38) · 0.25–1 kVA mosque (1) · 8 kVA school (1) · 6 kVA medical (1) · 4 kVA market (1) · 5 kVA PHASE BALANCE A/B/C: round-robin · imbalance < 7%
LV STEINER TREE · VOLTAGE DROP CONVERGENCE CUMULATIVE ΔU% PER FEEDER 5% limit 0m 240m 480m 0% worst 4.20% CONVERGED IN 3 ITERATIONS 20 poles · 19 edges · NFA2X 2×10mm² · ΔU max 4.20% · PUIL 2020 §3.16 OK
30-RULE IEC / PUIL VALIDATOR · LIVE CHECK ✓ V-001 Cumulative ΔU ≤ 5% (PUIL 2020 §3.16) 4.20% ✓ V-002 Cable CSA ≥ 1.5 mm² (SPLN 42) 4 mm² min ✓ V-003 Cable Iz ampacity (IEC 60364-5-52 §523) OK ✓ V-004 Cable thermal stress (§434.3) OK ✓ V-005 OCPD Icu ≥ Icc (IEC 60364-4-43) OK ✓ V-006 RCD 30mA on final circuits (PUIL §4.2.3) 20/20 OK ✓ V-007 OCPD selectivity (IEC 60898-1) OK ✓ V-008 Bus voltage ±10% (IEC 60038) 342–418 V ✓ V-009 Cable loading ≤ 100% Iz max 73% ⚠ V-010 Cable loading > 80% (warn) 1 cable @ 84% ✓ V-011 PV string Voc cold ≤ 1000 V (IEC 62548 §5.3) 812 V ✓ V-012 PV Vmpp hot ≥ inverter Vmin OK ✓ V-013 PV modules/string ∈ [Nmin, Nmax] 16/string ✓ V-014 BESS:PV ratio 1–5 kWh/kWp 2.6 kWh/kWp ✓ V-015 BESS:PV ratio in optimal band OK ✓ V-016 PE conductor on every cable (IEC 60364-6-61) OK ✓ V-017 Earthing electrode resistance ≤ 5 Ω 3.2 Ω ✓ V-018 SPD class T1+T2 on incoming OK ✓ V-019 Transformer rating headroom ≥ 20% 28% ✓ V-020 Transformer impedance vs Icc (IEC 60909) OK ✓ V-021..V-029 topology / catalogue / signatures 9/9 pass ✓ V-030 TCC selectivity ≥ 1.6:1 (IEC 60269) 39:1 ✓ V-031 PE-approval signature present HMAC-OK ✓ V-032 Twin provenance ≥ 50% 84% ✓ V-033 GOB confidence mean ≥ 0.65 0.82 ✓ V-034 Audit chain integrity OK 30 / 30 PASSED · 0 ERRORS · 1 WARNING · HMAC-SIGNED
DESA BELEME · 3D DIGITAL TWIN · INTERACTIVE 47 buildings · 124 kWp PV · 320 kWh BESS · 20 poles · NFA2X 2×10 cables Genesis run · 3D twin emitted · ready for Engineering Studio
Stage 1/7 · AOI Frozen Desa Beleme · -4.082, 138.624 Elapsed 0.18s
"Engineering is gated on a completed Genesis run. Twin provenance is stamped to every package — every pole, every cable, every meter linked back to a real WKT footprint and a real DEM elevation. The audit chain is the project." ADR-039 · Twin-First Pipeline
The Engine, Five Surfaces

Five surfaces.
One pass.

From the same coordinate, the engine produces five interlocking analyses: site bankability, grid intelligence, climate & hazard, political & stakeholder, and physical system design. Each surfaces as a structured report. Each is auditable. Each is generated in the same sub-two-minute pass.

"Does this village pencil?"
Layer 01 · Site Bankability

Composite financial picture, generated automatically

Internal rate of return, levelized cost of energy, sized capital stack, grant eligibility — computed from indexed inputs and rendered in DFI-readable form. The output is the same artifact a consultancy would deliver after twelve weeks, generated in one engine pass. Memoised < 50 ms even on 10k-village portfolio sweeps.

Twelve composite indices ranked side-by-side: economic, climate, political, regulatory, capital-pathway
Monte Carlo IRR distribution with explicit P50 / P90 confidence bands
Pre-filled grant eligibility checks against DAK Energi, GCF, USAID DCA, BPDLH, JETP — 12 named instruments
SAMPLE OUTPUT · Desa Beleme
Project IRR (P50)18.4%
LCOEUSD 0.18/kWh
Capital Stack60% grant / 30% debt / 10% eq
DAK Energi Eligible✓ Confirmed
MC P90 IRR12.1%
Carbon Revenue, 25-yrUSD 284K
Monte Carlo IRR distribution · 10,000 simulations
P50 18.4% P90 12.1%
"Will the grid arrive? When?"
Layer 02 · Grid Intelligence

Six grid-arrival signals, fused, with explicit disagreement

168 months of satellite nightlight trend, midline trace data at 250m, transformer geolocation, ESDM substation capacity, the full RUPTL Lisdes 2024–2033 extension pipeline, and live infrastructure deltas. Six independent sources fused into one composite grid-arrival probability. Disagreement flags surface uncertainty bands — never hidden.

Composite A-to-F grid-arrival grade with confidence intervals
Runway-before-grid estimation — the critical signal for minigrid vs grid-anticipation decisions
Explicit disagreement flags when sources contradict (e.g. RUPTL says 2030, nightlight extrapolation says 2042)
GRID SIGNAL COMPOSITE · Beleme
Grid Arrival GradeB+ (Moderate)
Nightlight Trend (168mo)Stable dark
RUPTL 2024–2033Not listed
MV Line Distance8.4 km
Runway Estimate12–18 years
ConfidenceHigh (5/6 signals)
168-month nightlight trace · flat-dark = no electrification
2012 2026 VIIRS DNB radiance · Beleme · 14yr flat
"What does climate cost this asset?"
Layer 03 · Climate & Hazard

9-hazard composite + CMIP6 ensemble + AAL reinsurance-calibrated

CMIP6 ensemble projections (4 SSP scenarios), 9-hazard composite (earthquake / volcanic / flood / tsunami / SLR / drought / extreme weather / landslide / wildfire), Swiss Re 2023 Asia-Pacific PV loss ratios, IEC TR 63157:2021 calibration. AAL %CAPEX, 25-yr expected loss USD, insurability grade A–F, TCFD/IFRS S2/ESRS E1 disclosure auto-fill.

Climate-accelerated battery cell degradation modeled over 25-year project lifetime
Complete climate adaptation file ready for insurance pricing at quoted 2.3% vs 5–8% generic band
TCFD-aligned physical risk disclosure ready for climate fund submission
CLIMATE RISK PROFILE
Annual Average Loss0.8% of CAPEX
Flood HazardLow (<0.2m, 5yr RP)
Compound Correlation0.34 (flood × slide)
Battery Degradation Δ+1.2%/yr (CMIP6 SSP2)
Insurance Rate2.3% (vs 5–8%)
TCFD StatusDisclosure Ready
9-hazard radar · composite score 72/100 · grade B+
QuakeVolcFloodTsunSLRDrghtStormLand
"What kills this project on the ground?"
Layer 04 · Political & Stakeholder

Live political risk intelligence · 22-step engagement playbook

Live ingestion of conflict events (ACLED), anti-corruption case tracking (KPK), electoral-window analysis (KPU Pilkada), adat territory overlap (BRWA 2,208 territories), pesantren density (Kemenag), governance percentile (WGI). 22-step engagement playbook with USD line items per stage, calendar months-to-COD, real contact list.

Pre-empts the most common cause of project failure in distributed energy
Pilkada electoral calendar integrated — avoids political interference windows
Insurance-formatted political risk narrative ready for MIGA/OPIC/NEXI submission
POLITICAL RISK SCORE
Composite ScoreB (Low-Moderate)
Conflict Events (12mo)0 within 50km
Corruption SignalNone flagged
Electoral Window2027 (3yr runway)
BRWA Adat OverlapOtsus eligible
PRI FormatReady
Stakeholder graph · 18 nodes · influence-weighted
VIL BupatiKadesPLNKyaiGubnBPD
"What gets built?"
Layer 05 · System Design

Hourly dispatch + 3D digital twin + per-HH engineering

Hourly solar-plus-storage dispatch under variable weather, climate-accelerated degradation curves, hybrid diesel-PV-BESS retrofit modeling, grid-adjacent sell-back analysis, and one-click 3D digital twin generation. Followed by full per-HH electrical design (see Engineering section).

PV array sized to terrain, ground coverage, and irradiance — not by rule of thumb
Complete LV network design routed to every building footprint from satellite data
Navigable 3D scene: draping terrain, extruded buildings on real WKT footprints, real river geometry
SYSTEM DESIGN OUTPUT
PV Array124 kWp (226 modules)
BESS320 kWh / 80 kW
Dispatch ModeSolar-first + peak shave
Annual Yield168,400 kWh
LV Poles20 (optimized)
3D TwinGenerated < 2 min
24-hour dispatch · PV (gold) + BESS (cyan) + load (white)
00h 12h 24h
PV Simulation Engine · ADR-043

Better than PVsyst.
In your browser.

A full physics chain rebuilt from first principles. 12-component loss stack. Hay-Davies POA. Sandia 4-rack thermal model. 72-bin per-village horizon shading (active, not baked-unused). Bifacial dynamic ROI/GCR. 1-axis tracking with backtracking. Indonesian field-calibrated degradation rates. ±1.5–2% browser accuracy vs pvlib reference, ±0.5–1% with Finkelstein-Schafer real TMY (server path).

FeaturePVsyst (de facto)Lumen8 (browser-native)Advantage
Transposition modelPerez (isotropic)Hay-Davies circumsolar + diffuse anisotropicTropics-tuned
Horizon shadingGeneric TMY bin72-bin per-village terrain profile (active, not baked-unused)+2–8% in highlands
Row shadingOptional addonGeometric bypass-diode modelIntegrated, free
Tracking (1-axis)Yes, with backtrackingYes, with backtrackingParity
Bifacial modelFixed +8–10%Dynamic ROI/GCR albedoSite-dependent
Sandia thermalFixed +30°CWind-speed dependent, 4 racking typesDynamic
Spectral AM correctionYes (generic)Kasten-Young + 5 module types (mono/poly/CdTe/HIT)Tropical-calibrated
Degradation rate0.7%/yr genericPurba 2020 Java field 0.68/yr primary + PID riskIndonesian field-sourced
BESS dispatchGeneric4-strategy: self-consumption / arbitrage / reserve / load-followingTariff-aware
String validatorBasicIEC 62548:2023 full (Voc cold ≤ 1000V)Procurement-safe
Bankability scorerNo6-component DFI score (P90/DSCR/PR/climate)Lender-ready
Per-HH engineeringNo (feeder-level)50 HH × (meter + RCBO + drop) per packageVillage-complete
Audit trailNoHMAC-SHA256 chain, twin-linkedBankable provenance
Offline-firstCloud / desktop installPWA + IndexedDB, 8,493 villages pre-bakedNo connectivity
Cost accuracyGeneric global9-region Indonesian pricingLocalised BOQ
Browser nativeNo (desktop app)Vanilla JS + Three.js, 2015 laptop OKZero install
12-Component Loss Waterfall
Nameplate STC → net energy yield · annotated with IEC clauses + Indonesian field calibration
Nameplate STC
100.0%
- LID year 1 (2.0%)
98.0%
- Spectral AM (0.5%)
97.5%
- Soiling (1.5%)
96.0%
- IAM incidence (3.0%)
93.0%
- Mismatch (2.0%)
91.0%
- DC cable (1.5%)
89.5%
- Connectors (1.0%)
88.5%
- Inverter eff. (3.0%)
85.5%
- AC cable (1.0%)
84.5%
- Transformer (1.0%)
83.5%
- Monitoring/parasit. (1.0%)
82.5%
- Climate degrad. (1.5%)
81.0%
NET energy yield
81.0%
12 user-configurable loss factors (was 4 hard-coded in legacy). Hay-Davies POA replaces Klein-Duffie isotropic (was ±3–5% error).
Sandia Tc replaces Faiman-lite (now ±0.5–1.5°C wind-sensitive).
Sources: Jordan et al. 2022 · Purba 2020 (Java field) · IEC 61853-2 · IEC 62548:2023 · IEC 61215.

±1.5–2% browser accuracy vs pvlib.
±0.5–1% server-side with FS real TMY.

The PV simulation engine is the result of a 10× physics overhaul (ADR-043) that replaced a 4-loss / Klein-Duffie / Faiman-flat / fixed-bifacial pipeline with a full 12-component / Hay-Davies / Sandia / dynamic-bifacial / 1-axis-tracking / 72-bin-horizon / Indonesian-degradation chain.

String validator catches procurement errors before BOM. BESS dispatch runs 4 strategies (self-consumption, flat-arbitrage, emergency-reserve, load-following) on a 25-yr SoH degradation curve. Bankability scorer surfaces a 6-component DFI grade.

ADR-043 · js/solar/pv-simulation.js (26 KB)
js/solar/bess-dispatch.js (10 KB)
scripts/bake-solar/server.py (pvlib + pvgis + NASA POWER)
Field calibration: Purba 2020 (5 Java sites, 5-yr) + Jordan 2022 meta-analysis
72-bin Per-Village Horizon Shading — active, not baked-unused

Most simulation tools store horizon profile metadata but never apply it — or use a generic flat-horizon assumption. Lumen8 bakes a 72-bin (5° azimuth) terrain profile per village from Copernicus DEM and applies it during the irradiance integration step. The yield delta is real: +2–8% in highlands, -1–4% in valleys with steep western horizons.

Desa Beleme
Papua Pegunungan · highland · 1,450m
+5.8%
vs flat horizon
Desa Topdan
NTT Timor · coastal flat · 80m
+0.3%
vs flat horizon
Desa Kelampai
Kalimantan · river delta · 12m
-1.2%
vs flat horizon
N E S W obstruction altitude (deg) Desa Beleme
Per-Household Engineering · ADR-038 / ADR-039

One coordinate. One engineering package.
PE-approval ready.

A full electrical system design built from the Genesis twin. 50 households, 50 meters, 50 RCBOs, 9 cables, 1 main MCCB. 30 IEC/PUIL validation rules. Dual-mode SLD (1-sheet A3 simplified or 6-sheet A0 full). TCC selectivity verified. Arc flash, harmonics, motor-start, IEC 60909 short-circuit. HMAC-SHA256 audit chain, 6-state PE-approval workflow, sealed construction-pack ZIP.

Single-Line Diagram · IEC 60617
Hover any element for IEC clause, rated current, cable CSA, voltage drop. Toggle simplified (1-sheet A3) vs full (6-sheet A0 with per-pole detail, phase balance, voltage profile).
PV 124 kWp 16str × 14mod INV 100 kVA BESS LFP 320 kWh / 80 kW MCCB 630A LV BUS · 400 V · 3P+N+PE F1·100A 14 HH ΔU 4.20% F2·80A 12 HH ΔU 3.80% F3·80A 12 HH ΔU 3.94% F4·80A 12 HH ΔU 4.18% TYPICAL HOUSEHOLD DETAIL · DESA-V-006 RCD ≤ 30mA · DESA-V-016 PE conductor FEEDER (NFA2X 2×16mm² Al) Pole-1 drop 17.8m M RCBO L1 MRCBOL2 MRCBOL3 MRCBOL4 MRCBOL5 MRCBOL6 + 44 more PE · 3.2 Ω
Element
Rating—
Bill of Materials · 50-HH Package
170-SKU catalogue with 9-region Indonesian pricing. Toggle between Java baseline and Papua Pegunungan (+15–40% logistics). 3 aggregations: flat / grouped / summary.
Total: USD 18,335 · 50 HH · Java baseline
QtySKUDescriptionUnit USDTotal USDStandardLead
50METER-1P-PREPAIDSingle-phase prepaid kWh meter, IP5442.002,100SPLN D3.0092 wk
50RCBO-30mA-16A-CRCBO Type C, 16A, 30mA, IEC 6100928.001,400IEC 61009-13 wk
50DB-IP44-4WAYDistribution board, IP44, 4-way35.001,750SPLN2 wk
50SPD-T3-1PSurge protection device, Type 3, single phase22.001,100IEC 61643-114 wk
1,012mNFA2X-2x10LV service drop ABC, 2×10mm² aluminium2.40/m2,429SNI 0420-3stocked
820mNFA2X-2x16LV feeder ABC, 2×16mm² aluminium3.10/m2,542SNI 0420-3stocked
180mNA2XSEY-3x35MV cable, 3×35mm² Cu, XLPE14.50/m2,610IEC 60502-25 wk
20POLE-9M-CONCConcrete pole, 9m, 200daN tip load82.001,640SPLN 304 wk
1MCCB-630A-3PMCCB main, 630A 3P+N, Icu 36kA, Curve C680.00680IEC 60947-24 wk
4MCB-100A-4P-CMCB feeder, 100A 4P, Curve C95.00380IEC 60898-13 wk
3MCB-80A-4P-CMCB feeder, 80A 4P, Curve C82.00246IEC 60898-13 wk
1EARTH-ROD-2.4M-CUCu earthing rod, 2.4m, 16mm dia38.0038IEC 62561-22 wk
4SPD-T1-T2-3PSPD T1+T2 combined, 3-phase, incoming285.001,140IEC 61643-115 wk
50 dLABOUR-INSTALLField labour: install + commissioning2.40/HH120PUIL 2020—
—SUBTOTAL EQUIPMENTMaterials only (Java)18,055
—SUBTOTAL LABOURField crew + commissioning280
—TOTAL (Java)50-HH packageUSD 18,335
TCC Selectivity Studio
Time-current curves for parent-child OCPD pairs, plotted log-log per IEC 60898 / 60947 analytic curves. Selectivity ratio annotated. Verified before construction pack ZIP issuance.
Current (A) — log scale Time (s) — log 10 100 1k 10k 36k 10000s 1000s 100s 10s 1s 0.01s RCBO 16A · IEC 61009 MCCB 630A · IEC 60947 Icc 28 kA SELECTIVE ratio 39:1 · ≥ 1.6:1

Selectivity Ratio Analysis

Parent: MCCB 630A · upstream
Child: RCBO 16A · 50 households
Icc fault: 28 kA at HH bus
✓ HIGHLY SELECTIVE
Ratio: 39:1
Required (IEC 60269): ≥ 1.6:1
51 parent–child pairs evaluated
100% selective across full Icc range
Fault on HH circuit isolates only that RCBO. Main MCCB unaffected. No cascade trip.
Arc Flash @ HH bus
Incident energy: 0.42 cal/cm²
PPE Cat 1 (per IEEE 1584-2018)
Working distance: 455mm
Label generator: NFPA 70E ready
Validation Engine · 30 Rules
DESA-V-001 .. DESA-V-034 — every rule cites a PUIL, IEC, IEEE, or SPLN clause. Average package: 1–2 informational findings, 0 errors. PE-approval gate halts on any error.
✓ 30 / 30 passed · 0 errors · 1 warn · HMAC-signed
✓
DESA-V-001
Cumulative ΔU ≤ 5%
PUIL 2020 §3.16 · 4.20% max
✓
DESA-V-002
Cable CSA ≥ 1.5 mm²
SPLN 42 · min 4 mm²
✓
DESA-V-003
Ampacity Iz
IEC 60364-5-52 §523
✓
DESA-V-004
Thermal stress
IEC 60364-4-43 §434.3
✓
DESA-V-005
OCPD Icu ≥ Icc
IEC 60364-4-43
✓
DESA-V-006
RCD ≤ 30 mA
PUIL §4.2.3 · 50/50 OK
✓
DESA-V-007
OCPD selectivity
IEC 60898-1 · 39:1
✓
DESA-V-008
Bus voltage ±10%
IEC 60038 · 342–418V
✓
DESA-V-009
Cable loading ≤ 100%
max 73% · headroom OK
⚠
DESA-V-010
Cable loading > 80%
1 cable @ 84% · advisory
✓
DESA-V-011
PV string Voc cold
IEC 62548 §5.3 · 812V
✓
DESA-V-012
PV Vmpp hot ≥ Vmin
inverter spec OK
✓
DESA-V-013
Modules/string in band
16 modules · in [12, 18]
✓
DESA-V-014
BESS:PV ratio
2.6 kWh/kWp · 1–5 OK
✓
DESA-V-015
BESS in optimal band
cycling pattern OK
✓
DESA-V-016
PE conductor on every cable
IEC 60364-6-61
✓
DESA-V-017
Earth electrode ≤ 5 Ω
IEC 62305-3 · 3.2 Ω
✓
DESA-V-018
SPD T1+T2 incoming
IEC 61643-11
✓
DESA-V-019
Transformer headroom ≥ 20%
28% · OK
✓
DESA-V-020
Tx impedance vs Icc
IEC 60909 · OK
✓
DESA-V-021
Topology connectivity
no orphan nodes
✓
DESA-V-022
Phase balance < 10%
round-robin · 7%
✓
DESA-V-023
Catalogue SKU integrity
170 SKUs match
✓
DESA-V-024
Arc flash < PPE Cat 4
IEEE 1584-2018 · Cat 1
✓
DESA-V-025
Harmonics TDD ≤ 5%
IEEE 519-2014 · 3.2%
✓
DESA-V-026
Motor start voltage dip
≤ 10% · pumping OK
✓
DESA-V-027
Short-circuit Icc
IEC 60909 · 28 kA
✓
DESA-V-028
Fuse-cable coordination
k²S² ≥ I²t
✓
DESA-V-029
PE-approval signature
HMAC-OK · audited
✓
DESA-V-030
TCC selectivity ≥ 1.6:1
IEC 60269 · 39:1
✓
DESA-V-031
LPS Class II
IEC 62305-2
✓
DESA-V-032
Twin provenance ≥ 50%
ADR-039 · 84%
✓
DESA-V-033
GOB confidence ≥ 0.65
mean 0.82
✓
DESA-V-034
Audit chain integrity
SHA-256 manifest OK
Construction Pack ZIP · Sealed Deliverable
10 directories, 26 files, 87 KB. HMAC-signed manifest. PE signature stamps the audit row. State-machine gate: must reach APPROVED before download. Once ISSUED, packs are immutable.
00_README/
1 file · 4.2 KB
README.txt · project meta + standards citations + audit hash
01_SLD/
6 SVGs · 487 KB · A0
sheet_01_overview.svg
sheet_02_feeder_F1.svg
sheet_03_feeder_F2.svg
sheet_04_earthing_LPS.svg
sheet_05_phase_distribution.svg
sheet_06_voltage_profile.svg
02_BOM/
3 XLSX · 28 KB
bom_flat.xlsx (per-HH detail rows)
bom_grouped.xlsx (by category)
bom_summary.xlsx (rolled-up totals)
03_Cables/
1 XLSX · 14 KB
cable_schedule.xlsx (IEC 61082-2 format)
04_Earthing/
2 files · 6 KB
electrode_test_plan.txt
lps_class_calculation.txt
05_Commissioning/
3 TXT · 8 KB
ir_test_checklist.txt
rcd_test_checklist.txt
commissioning_protocol.txt
06_AsBuilt/
1 CSV template · 3 KB
asbuilt_blank_pole_log.csv (50-row template)
08_Compliance/
2 files · 18 KB
validation_report.json (30 rules, signed)
jurisdiction_citations.md (PUIL/IEC refs)
09_PE/
1 JSON · 1.2 KB
pe_signature.json (actor + timestamp + HMAC)
manifest.json
SHA-256 manifest · pack_hash signed
SHA-256 of every file
pack_hash signed with project HMAC key
twin_run_id linked to Genesis run
PE-Approval State Machine · 6 States, HMAC-Chained
DRAFT
SUBMITTED
REVIEWED
APPROVED
ISSUED
AS-BUILT
Interactive 3D Digital Twin
Three.js scene with terrain mesh, extruded buildings on real WKT footprints, PV array, LV poles + drop cables. Layer toggles. Orbit + zoom + pan. Zero install — runs in any modern browser.
DESA BELEME · PAPUA PEGUNUNGAN · 47 BLDGS
Terrain
Buildings
PV Array
LV Poles
Hazards
47 buildings (GOB v3) 20 LV poles · Steiner 124 kWp PV · 226 modules 320 kWh BESS 72-bin horizon baked
Standards Stack PUIL 2020IEC 60364-5-52IEC 60898-1IEC 61009-1IEC 60909IEEE 1584-2018IEEE 519-2014IEC 62548:2023IEC 62305-2IEC 60617CGMES 2.4.15SPLN 42SNI 0420-3
Stage 4 · ADR-041 P1–P2

From PE-approved package.
To switch-on day.

Twelve construction-operations skills carry the audit chain forward from APPROVED → ISSUED → AS-BUILT. RFI management, daily reports, progress tracking, punch lists, procurement orders, safety checklists, quality audits, as-built logs — every action HMAC-signed and chained to the upstream Genesis + engineering audit row.

LIVE
Document Control
RFI Management
Auto-numbered RFI · urgency-driven deadline (low 14d / med 7d / high 3d) · photo ref · IEC/PUIL refs auto-populated · audit row signed
LIVE
Field Reporting
Daily Report
Timestamp · photo set · labour summary · safety incidents · material log · weather observations · GPS pin
LIVE
Analytics
Progress Tracking
% complete per milestone · earned value vs schedule baseline · crew productivity HH/day · burndown chart
LIVE
Closeout
Punch List
Open items · severity · owner · deadline · photo evidence · sign-off workflow · prioritized checklist
LIVE
Supply Chain
Procurement Order
Auto-match to BOM SKUs · vendor selection · lead-time estimate · payment schedule · TKDN compliance check
LIVE
QC
Safety Checklist
PUIL 2020 §3 electrical safety · grounding · PPE · isolation · LOTO · arc flash · PPE Cat 1 verified
LIVE
QC
Quality Audit
IR thermography checklist · loop impedance test plan · RCD trip-time verify · as-built photo walk-through
LIVE
Closeout
As-Built Log
Pole-by-pole GPS + photo · actual cable CSA (if differs from design) · HH meter readings · SN registry
LIVE
Resource
Contractor Profile
Crew composition · K3 electrical certification · prior projects · safety record · payment terms
LIVE
Risk
Contingency Log
Identified risks · mitigation budget tracking · weather delays · supply chain incidents · escalation
LIVE
Closeout
Warranty Tracking
Per-component expiry calendar · SN registry · vendor SLA · claim workflow · O&M handoff
LIVE
Analytics
Budget Status
Committed vs spent vs forecast · per-line variance · payment milestone tracker · cash flow projection
Genesis RunHMAC: 7a2f...
Engineering APPROVEDHMAC: e9d4...
RFI #001HMAC: 3b8c...
Daily Report 2026-05-04HMAC: f1a7...
Punch List #003HMAC: 9c5e...
As-Built UploadHMAC: 6d2b...
COD Sign-offHMAC: signed
"Every action HMAC-signed and chained from coordinate-zero to switch-on day. The audit trail is the project."
Procurement Intelligence

200 SKUs, sourced.
12 commodity risks, tracked.

A 200-SKU catalogue with TKDN/SNI vendor-scoring (sourced pricing 40 / standards 30 / source quality 20 / lead-time 10), six KAN-accredited certifier badges, copper / steel / silicon commodity-risk tracking, and a sealed-bid second-price marketplace for portfolio bundles. From BOQ to PO with provenance.

Catalogue Heatmap · 12 categories
PV modules
78
42 SKUs · 72% src · 6wk
Inverter
67
28 SKUs · 65% src · 5wk
BESS LFP
54
14 SKUs · 48% src · 10wk
LV cable (NFA2X)
84
18 SKUs · 92% src · stocked
Transformer
71
12 SKUs · 68% src · 6wk
MCB / RCBO
82
24 SKUs · 88% src · 3wk
MCCB
74
16 SKUs · 75% src · 4wk
SPD T1+T2
62
8 SKUs · 56% src · 5wk
Pole / racking
69
10 SKUs · 70% src · 4wk
Meter (prepaid)
79
9 SKUs · 82% src · 2wk
LPS / earthing
73
15 SKUs · 71% src · 3wk
Misc / fasteners
88
4 SKUs · 95% src · stocked
KAN-Accredited Certifiers (BSN scope)
Sucofindo
TKDN audit · K3 electrical · pre-COD inspection · ~2–4 wk
PT Surveyor Indonesia
SNI certification · government project preference · KAN-LP
LMK PLN
PLN technical compliance · SPLN standards · grid-tie cert
BKI
Marine + offshore + state asset compliance · KAN-LSPro
TÜV SÜD Indonesia
IEC 61215 / 61730 PV · IEC 62109 inverter · IEC 62133 BESS
SGS Indonesia
Module pre-shipment QA · cable type testing · global recognition
Sealed-Bid Second-Price Marketplace
Vickrey-Clarke-Groves auction logic. Vendors bid sealed on village bundles. Reveal happens at lockout. Winner pays the second-highest bid — the mechanism incentivises truthful pricing and prevents overbidding distortion. Funder transparency built in.
VENDOR
SEALED BID (USD)
RANK
Vendor C · PT Solusi Energi
$868,000 · #1 ✓
Vendor A · PT Energi Bersih
$924,000 · #2
Vendor B · PT Surya Andalan
$1,002,000 · #3
Vendor D · PT Cahaya Nusa
$1,108,000 · #4
Vendor E · PT Tenaga Mandiri
$1,212,000 · #5
✓ Vendor C wins. Pays $924,000 (Vendor A’s sealed bid, second price). Truthful bidding incentive preserved. 50-village bundle awarded, BOM auto-locked, pre-COD certifier (Sucofindo) auto-scheduled.
Copper LME spot
+15.2%
YoY · 2026 vs 2025 · cable cost +2.3% project CAPEX sensitivity · ECB / Frankfurter feed
Steel HRC
+8.4%
YoY · pole + racking exposure · World Bank Commodities feed
Silicon (PV grade)
-3.1%
YoY · downward pressure on module cost · BNEF spot
Grant Matching & Funding Stack

12 named windows.
$330M+ visible TAM.

A live grant scorer evaluates each village against twelve named instruments — WB EASE, ADB SRP, KfW DesaListrik, JICA, GEF, USAID DCA, APBN DAK Energi, UNDP, EU JTET, MIGA, Verra, JETP. Per-village fiscal-fit USD computed. Funding stack auto-assembled from Dana Desa + Otsus + DAK Energi + JETP + MDB catalog. Portfolio mode aggregates DFI submission-ready totals.

Desa Beleme Tier B · Highland $750k CAPEX WB EASE $50k · score 0.78 ADB SRP $40k · score 0.61 KfW DesaListrik $75k · score 0.83 JICA ENR $35k · score 0.55 GEF ACC $25k · score 0.49 USAID DCA $30k · score 0.52 APBN DAK Energi $200k · score 0.71 UNDP CRF $15k · score 0.41 EU JTET $40k · score 0.58 MIGA PRI 2.3% · 0.66 Verra VCS $25/tCO₂e · 0.62 JETP $150k · score 0.76
APBN DAK Energi Penugasan
Sponsor: Kementerian ESDM · Instrument: competitive grant · Otsus-compatible: yes
Annual envelope
IDR 680B (~USD 43M)
Per-project typical
USD 2M (range 0.5–6M)
Tier eligibility
Tertinggal · Sangat Tertinggal · Berkembang
Match-funding
10% kabupaten match required
Score for Beleme
0.71 (high)
Fiscal fit
USD 200k allocated
Dana Desa (5-yr stack)
$300k
Inpres 17/2025 · Kemenkeu DJPK 2026
Otsus (Papua eligible)
$25k
UU 11/2006 · UU 2/2021
DAK Energi Penugasan
$200k
ESDM IDR 680B annual envelope
JETP concessional
$150k
$21.7B portfolio commitment (≥$250k bundle)
KfW DesaListrik grant
$75k
Otsus-compatible bonus +0.08 score
ADB SRP concessional
$40k
Sustainable Rural Program
TOTAL ELIGIBLE FUNDING$790,000 · 105% of $750k CAPEX · 0% commercial debt

Dana Desa

band(IDM_status) × topoMult × papuaExtra(1.4×)

Universal village block grant (APBN). Per-village IDR 1.0–1.5B annual depending on IDM status band, topographical difficulty multiplier, and Papua region extra (1.4×).

Inpres 17/2025 · Kemenkeu DJPK 2026

Otsus (Special Autonomy)

if province ∈ [ACEH, 6× PAPUA]

Restricted to Aceh + 6 Papua provinces. Per-village USD 5–25k notional, drawn from provincial Otsus envelope (3–10% of provincial APBD). Hard-filter: throws error if non-eligible province.

UU 11/2006 (Aceh) · UU 2/2021 (Papua)

DAK Energi Penugasan

if IDM ∈ {Tertinggal, Sangat-T, Berkembang}

Tier-gated competitive grant. ESDM annual envelope IDR 680B (~USD 43M). Typical award USD 2M per kabupaten (range 0.5–6M). 10% kabupaten match-funding required.

ESDM Permen · DJK competitive process
AI Orchestration Layer

Multi-provider agentic intelligence.
RFP, vendor selection, build progress — automated.

A bounded agentic loop with 40+ MCP-shape tools. Multi-provider stack: Claude Sonnet/Opus/Haiku, OpenAI, NVIDIA NIM, Ollama (local). Per-conversation cost cap $0.50, daily cap $25. 13-adapter API fabric beneath. In-flight deduplication, 5-minute negative-cache backoff. RFP autogen, vendor scorecard, satellite build verification — all auditable, all cost-bounded.

Claude
Sonnet 4.5 · Opus 4.7 · Haiku 4.5
LIVE · primary
OpenAI
GPT-4o · GPT-4o-mini · o1
LIVE · fallback
NVIDIA NIM
Llama 3.3 · Nemotron · Mixtral
LIVE · enterprise
Ollama (local)
Llama 3.1 8B · Mistral 7B
LIVE · offline
village_search rank_portfolio run_genesis build_eng_pkg construction_pack funding_stack score_grants hazard_composite playbook_22step vendor_score generate_RFP Lumen8 Agent 10-iter bounded loop $0.50 / conv cap
Searchvillage_search · rank_portfolio · top-N · find-similar · filter-by-grade · query-by-grant · kabupaten-aggregate · provincial-rollup
8
Genesisrun_genesis · fetch_gob · classify_terrain · digitize_roads · route_lv · render_3d
6
Engineeringbuild_eng_pkg · generate_sld · run_validation · compute_tcc · arc_flash · construction_pack · generate_rfp · pe_approve
8
Economicscompute_irr · score_grants · build_funding_stack · monte_carlo · viability_gap · bankability_score
6
Climatehazard_composite · aal_curve · cmip6_delta · insurability_grade
4
Politicalplaybook_22step · stakeholder_graph · brwa_overlap · acled_proximity
4
Procurementcatalogue_search · vendor_score · commodity_risk · sealed_bid
4
Sample agent loop · canned animation

A real interaction: investor asks for Tier C villages in Papua with election runway and high solar resource. Agent spawns 4 tool calls, returns ranked top-5 + briefing PDF. Cost stamped. Audit row signed.

USER Find Tier C villages in Papua with election window > 18 months and GHI ≥ 5 kWh/m²/day. Rank by developability, build funding stack on top 5, generate briefing PDF.
Iterations: 0/10 Tokens: 0 Cost: $0.00 / $0.50 cap Time: 0.0s Status: idle
Satellite-Verified Build Tracking

Sentinel-2 + YOLOv8.
Build progress, automated.

Continuous IoT telemetry plus quarterly Sentinel-2 satellite verification. YOLOv8-m detection model trained on 4k positives + 8k negatives identifies rooftops, access roads, PV arrays, and BESS containers. Confidence-thresholded matches generate Verra-compatible attestations. The MRV evidence chain extends from coordinate-zero to as-built.

Sentinel-2 L2A · 2026-04-12 · cloud 8% YOLOv8-m · conf 0.94 · matched BLM-12 · Δ 4.2m Stage: PV array online
2025-Q1 · bareQ2 · accessQ3 · foundationsQ4 · PV array2026-Q2 · complete
YOLOv8 detection confidence over 6-month build window
0.42 0.51 0.72 0.89 0.94 0.96
MRV evidence pipeline · coordinate-zero to attestation
Sentinel-2 STAC pull
cloud filter <20%
YOLOv8 inference
match-to-design
confidence threshold
audit-chain row
Verra attestation
Transparency · Phase 3 status: The detection model architecture, training set (4k positives + 8k negatives), and Planetary Computer STAC integration are production-ready. Live inference rolls Q3 2026 once the YOLOv8-m training run completes. Synthetic detections (160 confirmed + 25 review-queue) are in place for pipeline validation today. Architecture is real. Production inference: in flight.
Climate & Physical Hazard

9 hazards. Composite to insurability.
25-year CMIP6 ensemble.

Every village scored against a 9-hazard composite (earthquake / volcanic / flood / tsunami / SLR / drought / extreme weather / landslide / wildfire). Annual Average Loss calibrated to Swiss Re 2023 Asia-Pacific PV loss ratios. Insurability grade A–F. TCFD / IFRS S2 / ESRS E1 disclosure auto-fill. CMIP6 ensemble 4 SSP scenarios for 25-year asset planning.

Earthquake Volcanic Flood Tsunami SLR Drought Storm Landslide 100755025 72 grade B+
Annualised Average Loss (AAL) · %CAPEX
0.8% / yr25-yr expected loss · USD 78,400
B+
Insurability: Standard market Premium 2.3% vs 5–8% generic band · Swiss Re 2023 calibration · IEC TR 63157:2021
✓ TCFD / IFRS S2 / ESRS E1 disclosure auto-fill ready
CMIP6 SSP2 25-yr trend · Beleme
Temp +1.4°C (vs 2025)
Precip -3.2% / yr
Solar yield -0.8% / yr
Wind +4.1% peak
Sources: BNPB InaRISK · USGS · Smithsonian GVP · NASA SLR · Copernicus CMIP6 · Aqueduct · Think Hazard
Methodology: IEC TR 63157:2021 + Swiss Re 2023 Asia-Pacific PV loss ratios (0.08–0.53%/yr per hazard) + ADB TA-9150-INO 2024 CAPEX benchmark
Political & Stakeholder Intelligence

22-step engagement playbook.
USD line items per stage.

Live ingestion of conflict events (ACLED), anti-corruption case tracking (KPK), electoral-window analysis (KPU Pilkada), 2,208 BRWA adat territories, pesantren density (Kemenag), governance percentile (WGI). 22-step engagement playbook with USD budget per stage, calendar months-to-COD, real contact list. Insurance-formatted political risk narrative for MIGA / OPIC / NEXI submission.

Stage 1M 0–1
Scoping
Initial outreach to provincial / kabupaten authorities, PLN technical lead, ESDM regional. Project framing, data-sharing MOU, initial site visit clearance.
Bupati · Kadis ESDM · PLN distributing
3 steps · always required
$5.5k3 actors × USD 1.8k avg
Stage 2M 1–3
Community Entry
FPIC if BRWA adat overlap. Religious networks: PCNU (NU) / PDM (Muhammadiyah) / Church / adat council. Village officials: Kades, BPD. Conditional on village religious + adat profile.
Adat FPIC · PCNU/PDM · Church · Kades
5–7 steps · conditional
$25–55kFPIC alone $25k if required
Stage 3M 3–9
Formal Approvals
AMDAL (EIA), BKPM OSS license, BPN land certificate, KLHK IPPKH (forest clearance if applicable), provincial perizinan. Regulatory critical path — longest single stage.
AMDAL · BKPM · BPN · KLHK · perizinan
5 steps · regulatory
$55k+permits + studies + filings
Stage 4M 9–18
MOU / Close
PPA negotiation with PLN or off-taker, EPC contract sign, lender due diligence (DFI / commercial), insurance binding, escrow setup.
PPA · EPC · lender DD · insurance
4–6 steps · always
$65k+legal + DD + advisory
Stage 5M 16+
Construction & COD
Grievance Redress Mechanism (GRM) live, media programme, COD inauguration ceremony, handoff to O&M. Audit chain transitions to Stage-4 ops cockpit.
GRM · media · COD ceremony · O&M
3 steps · always
$43kcommunity + ceremony + media
Total playbook budget
USD 200–300k
base + (adat FPIC × $25k) + (forest IPPKH × $20k) + (election × $15k)
Months to COD
18 + δ
δ = (warning-flags × 2) + (election-window × 4) + (BRWA × 3)
Stakeholder Graph · 18 nodes · influence-weighted
Force-directed layout. Edge weight from path-finding algorithm. Hover any node to highlight its influence chain. Sources: BRWA (2,208 adat) · KPA CATAHU · KPU Pilkada · KPK · WGI · ACLED · Walhi/LBH · Kemenag pesantren registry · AMAN.
Village Bupati PLN Kades Adat Gubernur Camat BPD ESDM BUMDes PCNU Kyai Walhi KPK KPA PT SMI BPDLH
Audit & Evidence Chain

Every metric.
Sourced, signed, auditable in both directions.

Each output ships with a Merkle-tree-style HMAC-SHA256 evidence chain. A reviewer audits the chain, not the conclusion. Click any metric to surface every contributing source, vintage, confidence interval, and disagreement flag. The build-status registry tracks 156 components — LIVE / BUILT / STUBBED / SPECIFIED / BLOCKED / VERIFY — in real time.

Project Approval7a2f9c4e1b5d8... Engineering Pkge9d4f6a2... Economics Profile3b8c1f5e... Political Scoref1a72d8b... Climate Profile9c5e3d72... SLD svg8b4a.. BOM xlsx2f7c.. val.jsond1e9.. funding6a3b.. ATP/WTPc5f8.. CMIP69e2d.. AAL4b7a.. GOB v3conf 0.82 170 SKUJava pricing 12 grantsInpres 17/2025 Numata 2022Sumba DCE CopernicusSSP2-4.5 EVERY NODE = HMAC-SHA256 HASH · EVERY EDGE = SIGNED RELATIONSHIP tamper-evident · append-only · bidirectional trace · Article 6.4 ready Click any metric on the right →
Click-to-trace explorer · IRR P50 = 18.4%
IRR P50 18.4% — derived
↑ Capital stack + LCOE — funding-stack.js
↑ Funding stack 12 instruments scored
↑ BOM $18,335 (Java baseline)
↑ PV yield 168.4 MWh/yr
↑ PV simulation ±1.5–2% vs pvlib
↑ Hay-Davies POA + Sandia Tc + 72-bin horizon
↑ Sentinel-2 + Copernicus DEM + village 8493-baseline
Build-status registry · 156 components
62
LIVE
38
BUILT
12
STUBBED
32
SPECIFIED
4
VERIFY
8
BLOCKED
Weekly burn-down · unshipped components
76 target 0
Disagreement flag · surfaced not hidden
Grid arrival year:
• RUPTL Lisdes 2030 (PLN official)
• Nightlight extrapolation 2042 (VIIRS DNB)
Δ = 12 years uncertainty
Composite reflects band, not point estimate.
Report Engine · 10 Formats

Ten artifacts.
Generated, not template-filled.

Each artifact is dynamically generated from live engine outputs, not template-filled, and each is calibrated to the institution that will read it. One engine pass, ten different rendered formats. Hover any card for sample first-page preview.

For Governments
Ministerial Briefing
2 pages · cabinet-grade
Cabinet-grade summary with national portfolio map and headline impact metrics. Print-ready.
Kementerian ESDM · BAPPENAS · Kemendes PDT
RUPTL Lisdes Briefing
Cabinet — 2026 Q2
Pipeline Overview
Top 50 Villages
Funding Mix
Political Assessment
Stakeholder map + 22-step
Coalition path, electoral calendar, anti-corruption exposure score, 22-step coalition playbook with USD line items.
BAPPENAS · Kemendes PDT
Political Risk Assessment
Desa Beleme — Papua Pegunungan
Stakeholder Graph
22-Step Playbook
ACLED + KPK signals
Quarterly Portfolio
Portfolio KPIs + burndown
Pipeline KPIs, deployment velocity, capital committed, households electrified.
Internal · LP Reporting
Quarterly Portfolio
Q1 2026
Pipeline Velocity
Capital Deployed
Households Electrified
For Climate Funds
TCFD Climate Report
Physical risk disclosure
CMIP6-driven physical risk disclosure aligned to TCFD recommendations. IFRS S2 + ESRS E1 fields auto-fill.
GCF · GEF · Climate Funds
TCFD Climate Disclosure
Beleme · Papua
Governance
Strategy
Risk Management
Metrics & Targets
Grant Application Draft
12 instruments · pre-filled
DAK Energi, USAID DCA, GEF, Green Climate Fund, KfW DesaListrik, ADB SRP, JICA — pre-filled, ready to submit.
GCF · USAID · GEF · DAK
DAK Energi Application
Tier C · Papua eligibility
Project Description
Budget Breakdown
Match Funding
MRV Annual Report
Sentinel-2 + IoT verified
Delivered impact verification, Verra and Gold Standard compatible, Article 6.4-ready evidence chain.
Verra · Gold Standard · Art. 6
Annual MRV Attestation
2026 — Verra VCS
Baseline
Delivery (Sentinel-2)
Audit Chain
For DFI Underwriters
Bankable Project Report
63 pages · IFC PS 1–8
Full financial model, capital stack, sensitivity analysis, risk register. Auto-generated in ~90 sec, browser-side jsPDF.
IFC · ADB · KfW · JICA
Bankable Project Report
Beleme · 50 HH · USD 750k
Executive Summary
Financial Model
Risk Register
ESAP Commitments
DFI E&S Assessment
IFC PS 1–8 checklist
Land rights status, environmental triggers, E&S pre-screening for portfolio and project level.
World Bank · IFC · FMO
E&S Assessment
IFC PS 1–8
PS1 Assessment
PS5 Land Acquisition
PS7 Indigenous (FPIC)
ESAP
PRI Submission Pack
Insurance-ready · 2.3% rate
MIGA, OPIC, and Nippon Export underwriter-specific political risk narratives.
MIGA · OPIC · NEXI
PRI Submission · MIGA
Indonesia · Beleme
PSN Risk Score
Stakeholder Matrix
ACLED Proximity
Premium Rate 2.3%
For Field Teams
Field Briefing
Offline-ready A4
GPS coordinates, road access, logistics checklist, local contacts for EPC and field teams.
EPC · Field Teams
Field Briefing
Beleme · Papua
Coordinates · Access
Local Contacts
Logistics
Who Lumen8 Is Built For

Three buyers. Three willingness-to-pay curves.
One platform.

The same engine produces three different strategic outcomes depending on who is holding it. Developers run thousands of feasibility studies in the time it used to take to scope one. Governments turn aspirational masterplans into ranked, financeable pipelines. Capital licenses the engine as its internal screening and verification surface.

01 · Developers
Run a thousand feasibility studies a week.

Conventional business development demands an USD 80,000–150,000 feasibility study before any financier will entertain a project. For sub-USD 1 million distributed-energy sites, this kills the business case before it begins. Lumen8 collapses BD cost to the marginal compute of one platform run. A developer can screen 1,000 candidate villages in a week, advance the top 50 to portfolio aggregation, and present a USD 25M bankable SPV in the time it used to take to scope a single feasibility study.

Pipeline conversion from 35% industry baseline to 70–80% projected. BD cost-per-MW down by an order of magnitude.
02 · Government
The country's rural electrification intelligence layer.

Every electrification masterplan in Indonesia — ESDM, BAPPENAS, PLN, Kemendes PDT — lists hundreds of pending sites. None are ranked by bankability. None are formatted for the financiers that would fund them. Lumen8 is not a vendor service to government. It is the country's view of its own rural electrification pipeline. Ingest the masterplan, score every site, output a five-year deployment sequence calibrated to the funding mix the country can actually mobilize.

A 30-second portfolio dashboard that survives a Cabinet briefing. A five-year tranche PDF that survives a World Bank credit committee.
03 · Capital
The screening and verification surface for every dollar going in.

Every financier — DFI, climate fund, political risk insurer — is constrained by the same bottleneck: project-preparation throughput. Lumen8 licenses to these institutions as their internal pipeline-screening, due-diligence-acceleration, and MRV-verification surface. Outputs are pre-aligned to IFC, ADB, KfW, MIGA, GCF, and Verra formats. A six-month due-diligence cycle compresses to four weeks. A USD 1.5M PPF deploys 3–5× further.

The standard becomes the format. The format becomes the procurement requirement.
The Commercial Unlock

Sub-USD 1M sites. Aggregated to
USD 25M tickets. Underwritten in weeks.

Financiers require minimum tickets of USD 10–25 million. Distributed-energy sites are USD 0.4–1 million. Without aggregation, no underwriter can deploy. Without per-site preparation, no aggregation is possible. Lumen8 closes this loop — and the artifacts the engine generates pass due diligence on every leg of the capital stack.

Grants & viability gap
15–25%
GCF · BPDLH · DAK Energi · Dana Desa
Concessional debt
30–40%
FMO · DFC · ADB · KfW DEG · JICA
Commercial debt
20–30%
ING · Rabobank · MUFG (with MIGA cover)
Sponsor equity
15–20%
Co-developer + EPC
Carbon pre-purchase
5–10%
Verra · Gold Standard · Art. 6.2 / 6.4

Aggregation that survives DD

A 50-village portfolio at USD 500K per site = USD 25M ticket. Lumen8 prepares all 50 sites in three days, generates a single SPV-level Bankable Project Report, and outputs IFC-grade documentation at portfolio level. The same DD pack covers concessional debt, commercial debt, and political risk insurance.

Multi-AE pathway

Each Lumen8 site comes pre-formatted for the Accredited Entity that suits it. PT SMI for sovereign-flavored programs. FMO for Dutch-aligned commercial programs. KfW for German bilateral. ADB for sovereign loans. Replicable structures, not one-off engagements.

GCF PPF compatibility

Lumen8's outputs satisfy GCF Project Preparation Facility cost categories: feasibility studies, ESIA, gender assessment, climate rationale, pre-FEED engineering, financial modeling, stakeholder consultation. A USD 1.5M PPF deploys 3–5× further when prep work is platform-generated.

Every Site ≡ A Verifiable Credit

Carbon, by design.
Not as an afterthought.

Every village in the Lumen8 dataset is documented, baselined, and instrumented to generate verifiable digital-MRV credits under Verra's emerging methodologies, Gold Standard for the Global Goals, and Article 6.4 of the Paris Agreement. This is not bolt-on accounting — it is engineered into the dataset itself.

01
Additionality, demonstrated
12–18 year grid runway estimate per village, derived from six independent signals (RUPTL, nightlight, MV distance, transformer geo, ESDM substations, infrastructure deltas). The platform demonstrates — not asserts — that without intervention, no electrification was forthcoming.
02
Baseline, documented
Pre-project household kerosene, candle, micro-diesel consumption profiles for every site. ATP/WTP from Numata 2022 Sumba DCE + BPS SUSENAS 2022. Each baseline shipped with HMAC-signed evidence trail to upstream sources.
03
Delivery, verified
Continuous IoT telemetry plus quarterly Sentinel-2 verification (YOLOv8 detection, nightlight delta, building electrification signature). Methodology-agnostic, registry-ready: Verra, Gold Standard, Article 6.4.
A Lumen8-prepared site is, on day zero, a digital-MRV-grade primary credit issuer. The platforms that lock the format become the standard. Carbon is not a side-revenue — it is a structural unlock for the capital stack.
Built for procurement, not retrofitted to it

Sovereign-grade compliance.
Foundations, not features.

Lumen8 is engineered for the procurement realities of the institutions that will use it. Indonesia's data residency requirements (UU 27/2022 PDP), EU GDPR for Dutch entity operations, and the auditability standards that DFIs demand — all foundational, not bolt-on.

Compliance Foundations

✓
Indonesia data residency. UU 27/2022 PDP-aligned. On-premises deployment available. Sensitive datasets stay in-country.
✓
EU GDPR. Dutch entity, EU operational base. DPAs and SCCs available for cross-border processing.
✓
Role-based access. Government-tier and institution-tier deployments with audit logs and SSO integration.
✓
SOC 2 roadmap. Type I targeted Q3 2026; Type II by Q1 2027.
✓
Knowledge graph. 15,941 nodes · 47,621 edges · 267 communities · GitNexus + graphify dual indexing.

Data Provenance

✓
Tamper-evident audit log. HMAC-SHA256 hash-chained, append-only. Every output carries a verifiable evidence chain.
✓
Free, Prior, Informed Consent. Built-in FPIC workflow with stakeholder mapping and consultation tracking.
✓
Source attribution. Every metric maps back to its dataset, vintage, and confidence interval. No black-box outputs.
✓
IFC PS 1–8 alignment. Performance Standards integrated into E&S pre-screen output.
✓
Digital MRV-grade. Article 6 Supervisory Body and ICVCM Core Carbon Principle compatibility.
A Lumen8 output is auditable in both directions. A reviewer can trace any conclusion back to the contributing source, timestamp, and confidence interval — and, conversely, can challenge any single source and see exactly which downstream conclusions are affected. That is the difference between a generated report and a typical consultancy deliverable.
From Indonesia, to Adapter Framework

Indonesia is the proof.
Not the market.

8,493 villages, fully indexed and operational, is the hardest country to do first — thousands of inhabited islands, four-tuple administrative hierarchy, sovereign data-residency rules, multi-source grid intelligence. Once the adapter framework is built, the Philippines, Papua New Guinea, Vietnam, Nigeria, and Kenya are months — not years.

Vector 01

Country expansion

The Indonesia engine is itself an adapter framework: replace the spatial index, swap the climate ensemble, plug in the local grid masterplan. Six target geographies pre-modeled.

PHL · VNM · PNG · NGA · KEN · ZAF · CMR
Vector 02

Vertical expansion

The same intelligence substrate generalizes to adjacent infrastructure types: rural water, telecommunications backhaul, agri-cold-chain, micro-mobility. Same engine. Same evidence layer. Different output schema.

Water · Telecom · Cold-chain · Mobility
Vector 03

Standard-setting

Whoever locks the digital-MRV format and the bankability evidence schema becomes the procurement standard. The window is the next 24 months.

Format → Standard → Procurement
CountryStatusUnelectrified SitesCapital Pool
Indonesia● Operational8,493 villagesBPDLH, DAK Energi, GCF, ADB, JETP $21.7B
Philippines● Engineering~2,400 sitiosNEA, DOE, ADB, JICA, USAID
Papua New Guinea● Engineering~6,500 villagesADB, World Bank, AusAid, NZAID
Vietnam● Scoping~1,200 hamletsEVN, JICA, KfW, IFC
Nigeria● Scoping~22,000 communitiesNEP, REA, AfDB, World Bank, GCF
Kenya● Scoping~3,800 sub-locationsREREC, KfW, AfD, IFC, USAID
South Africa● Scoping~1,500 villagesDBSA, IDC, GCF, KfW
Three Convergent Windows

Why this. Why now.

Three independent windows are converging at the same time. Sovereign demand, capital deployment pressure, and standard-setting all peak in the same 24-month corridor. The platforms that occupy this category by 2027 will be the procurement default for a decade.

01
Sovereign demand window
Indonesia's RUPTL Lisdes 2024–2033 mid-term review creates immediate institutional appetite for ranked, fundable rural pipelines. Multiple ministries are simultaneously evaluating intelligence platforms for portfolio-level deployment. JETP's $21.7B partnership commits real capital. First-mover access is now.
02
Capital deployment window
Concessional capital pools — GCF, GCF PPF, DFI bilateral programs — are under 2026–2027 deployment pressure with insufficient bankable artifacts to absorb. Lumen8 is the throughput unlock. Pipelines that exist on the platform absorb capital that would otherwise sit idle.
03
Standard-setting window
Article 6.4 Supervisory Body and ICVCM are converging on digital-MRV standards. Every village onboarded under platform-grade evidence today is a primary credit two years out. The platforms that lock the format become the standard. Whoever sits on top of this evidence layer reads from the source of truth for a decade.
Four Strategic Narratives

The same engine, four legitimate stories.
Each addressed to a different reader.

Lumen8 is a deep-tech engine, a financial-aggregation surface, a sovereign-electrification intelligence layer, and a carbon issuance primary — depending on who is holding it. Each narrative is true. Each reader cares about a different one.

N01 · To investors
Vertically integrated category-defining play
Three layers fused into a single product: the dataset (Indonesia at GOB v3 building resolution), the engine (Genesis → engineering → operations), the evidence layer (HMAC-signed Merkle audit). The compounding moat is two years on the dataset, four on engine + evidence in combination, decisive once standard-setting locks in.
Two-year dataset moat · multi-vector defensibility
N02 · To financiers
A USD 100B aggregation unlock for distributed energy
Concessional capital is structurally constrained by the per-site preparation arithmetic. Lumen8 collapses that arithmetic and prepares portfolio-grade artifacts that pass IFC, ADB, KfW, MIGA, GCF due diligence at the speed of platform compute. The bottleneck moves from preparation to underwriter throughput.
USD 100B+ capital unlock · 3–5× PPF efficiency
N03 · To governments
Sovereign electrification intelligence layer
A national rural electrification masterplan is, today, a list of names without ranks. With Lumen8 it becomes a ranked, financeable, schedulable deployment sequence calibrated to the funding mix the country can mobilize. The country owns its own pipeline.
Cabinet-grade dashboard · tranche-ready deployment
N04 · To climate funds
Article 6.4-grade primary credit issuer at scale
Every site documented, baselined, and instrumented to digital-MRV standards on day one. Verra, Gold Standard, Article 6.2 / 6.4 compatibility built into the dataset architecture. The platforms that lock the format become the standard.
Day-zero MRV-grade · multi-registry compatible
Capital Tranches

From operational engine.
To category default.

Five capital tranches across thirty-six months. Each tranche unlocks a discrete capability and de-risks the next. Tranche 1 is operational today. Tranche 2 is in delivery. Tranches 3–5 are sequenced to scale alongside customer onboarding.

T1
Tranche 01Indonesia EngineOperational
8,493 villages indexed. Genesis Pipeline live. Engineering Studio (30 IEC/PUIL rules, dual-mode SLD, TCC, arc flash, construction pack). Ten report formats generating. Engine compute is operational on production hardware. Status: deployed.
T2
Tranche 02Government Pilot & First ClosesIn Delivery
Multi-ministry pilot deployment. First USD 25M aggregated SPV close. First DFI institutional license. First climate-fund grant submission generated end-to-end on platform. Stage-4 construction operations cockpit live. Unlocks: live customer revenue, sovereign reference, and DFI procurement footprint.
T3
Tranche 03Adapter Framework + Two Country ReplicationsCommitted
Spatial index, climate ensemble, grid-intelligence layer abstracted into a country-replication framework. Philippines and Papua New Guinea brought to operational state. Multi-tenant infrastructure (Celery + Redis) scales engineering kernel. Unlocks: SaaS multi-country revenue and adapter-as-product economics.
T4
Tranche 04MRV-as-a-service & Carbon IssuanceRoadmap
Continuous IoT + Sentinel-2 + YOLOv8 satellite verification stood up at portfolio scale. Verra, Gold Standard, Article 6.4 issuance pipelines instrumented. Lumen8 becomes a primary-credit issuer for partner project portfolios. Unlocks: high-margin recurring revenue and standard-setting traction.
T5
Tranche 05Vertical Expansion: Water, Telecom, Cold-chainRoadmap
Same intelligence substrate generalized to adjacent rural infrastructure verticals. Output schemas and partner integrations extended. Unlocks: TAM expansion from rural energy to rural infrastructure as a category.
Hard Questions, Direct Answers

The objections we expect.
Answered, on the record.

The platform-grade questions we hear from analysts, underwriters, ministers, and engineers. Direct answers, no marketing veneer.

GIS visualizes existing data on a map. CAD tools design a single site to engineering specification. PVsyst simulates a known PV system. Lumen8 generates an entire feasibility study — economic, climate, political, system-design, regulatory, plus a complete per-HH electrical engineering package with 30 IEC/PUIL validation rules — across thousands of sites in parallel, ships it in ten different DFI-aligned formats, and carries an HMAC-signed audit chain back to source. GIS visualization and CAD design are surfaces of the engine, not the engine. Comparing Lumen8 to either is a category error.
Three-layer composition. The dataset alone is the moat for two years — Indonesia at 8,493-village resolution joined to GOB v3 footprints, BNPB hazards, ACLED conflict events, BRWA adat, KPU Pilkada, RUPTL Lisdes, and 168 months of nightlight is operationally hard to assemble even with capital, given multi-ministry data-sharing requirements, sovereign data residency (UU 27/2022 PDP), and the integration work. The engine and the evidence layer extend the window to four years. Replicating across countries does not multiply linearly — each country has its own data-residency rules, its own grid-intelligence sources, its own political topology. The compounding moat is what defines the category.
Every output ships with an HMAC-signed Merkle-style evidence trail showing every contributing source, vintage, confidence interval, and disagreement flag between contradictory inputs. The reviewer audits the chain, not the conclusion. This is faster and more rigorous than auditing a human-authored consultancy report — where the analyst's reasoning is opaque, the underlying sources are typically not enumerated, and confidence intervals are absent. The engineering package itself enforces 30 named PUIL/IEC validation rules and a 6-state PE-approval workflow before any construction-pack ZIP is issued.
Browser path: ±1.5–2% vs pvlib reference. Server path with Finkelstein-Schafer real TMY: ±0.5–1%. The 12-component loss stack (Hay-Davies POA, Sandia 4-rack thermal, Kasten-Young spectral AM, dynamic bifacial ROI/GCR, geometric row shading, 1-axis tracking + backtracking, 72-bin per-village horizon shading) replaces a legacy 4-loss isotropic model. Degradation calibrated to Purba 2020 Java field data (0.68%/yr primary, PID risk +0.15%/yr unmitigated). Source: ADR-043, Jordan et al. 2022 meta-analysis, IEC 61853-2.
Every construction-pack ZIP is HMAC-SHA256-signed, twin-linked (every cable and pole traceable to a real WKT footprint and DEM elevation), and gated on a 6-state PE-approval workflow (DRAFT → SUBMITTED → REVIEWED → APPROVED → ISSUED → AS-BUILT). Inside the pack: 6-sheet IEC-60617 SLDs, BOM with 9-region pricing and IEC/SPLN standard citations, cable schedule (IEC 61082-2), earthing test plan (IEC 62305), commissioning checklists, validation_report.json with all 30 named rules signed, jurisdiction citations (PUIL 2020 §3.16, etc.), pe_signature.json with actor + timestamp + HMAC. The audit chain is the project.
The 13-adapter API fabric is built for resilience: IndexedDB persistent cache, in-flight deduplication (6 callers → 1 API hit), 5-minute negative-cache backoff after 3 consecutive failures, fallback adapter chain (e.g. Sentinel-2 fallback to Esri imagery, GOB fallback to OSM building footprints, BNPB fallback to USGS + Smithsonian + NASA SLR + Aqueduct individually). The 8,493-village baseline is pre-baked at 50.9 GB. A laptop in the field with no connectivity can still run the full engine on any of those villages — Genesis, engineering, economics, BOM, validation, audit chain, construction-pack ZIP. Offline-first is not a feature, it is the architecture.
On-premises deployment available for ministerial / BUMN customers (PLN, ESDM, BAPPENAS, Kemendes PDT). Sensitive datasets stay in-country — the engine runs against an on-prem Postgres + Celery + Redis stack with the same browser surface. Cross-border processing for non-sensitive data uses a Dutch-entity EU operational base under GDPR with DPA / SCC paperwork. Role-based access for government tier (read-only, sovereign masterplan) vs institution tier (full engineering + procurement). Audit logs are signed and tamper-evident. SOC 2 Type I targeted Q3 2026, Type II by Q1 2027.
The Window is Now

The engine is operational.
The dataset is live.

Three buyers. One platform. Indonesia today, the category by 2027.

For Developers

Pilot pricing

USD 5K project credits + portfolio licensing. First five partners get founding-customer access.

Run the engine →
For Government

Sovereign engagement

Multi-ministry briefing pack. Pilot deployment within 60 days. Indonesia data residency.

Run the engine →
For Capital

Institutional license

Underwriter-tier access. PPF efficiency demo. DD-acceleration onboarding.

Run the engine →
Operational today
8,493 villages indexed
30 IEC/PUIL rules
10 DFI report formats
HMAC audit chain
Sovereign-grade compliance