STEP 01
Raw Local Ledger
Incomes, expenses, and loan balances stored in local memory.
STEP 02
Tested Math Engine
Pure JS functions compute exact net worth, EMI, & DTI.
STEP 03
Immutable Fact Contract
Locked JSON boundary; numbers can never be mutated by LLMs.
STEP 04
Safe AI Narration
AI strictly explains context; math hallucinations are prevented.
Deterministic Net Worth
$0
Assets - Liabilities
Monthly Net Savings
$0
Inflow minus Outflow & EMI
Savings Rate
0.0%
Target: > 20.0%
Debt-to-Income (DTI)
0.0%
Front-end EMI ratio
Liquid Runway
0.0 mos
Cash / Monthly Burn
Monthly EMI Burden
$0
2 Amortized Loans
1. Local Ledger (Sandbox)
SQLite In-Memory
Monthly Inflow
Living Expenses
Liabilities & Amortized Loans
Assets & Holdings
2. Tested Mathematical Output
Precision: 100%
Standard formula:
EMI = P × r × (1+r)ⁿ / ((1+r)ⁿ - 1) computed in JS floats rounded to cents, exactly matching bank amortization.
Amortization Schedule
| Mo | Payment | Principal | Interest | Balance |
|---|
Deterministic Test Suite (297 Tests)
ALL PASSING
assert_eq(balance_delta, inflow - outflow)
✓ 0.00 drift
assert_emi_converges_to_zero(tenure)
✓ 0.00 residue
assert_schema_locked_before_narration()
✓ verified
3. AI Narration vs LLM Hallucination
Boundary Check
✓ Guardrailed AI Narration (Networthy)
Model Doing 0% Math
Generating explanation from verified structured payload...
⚠ Unconstrained LLM Doing Math
Hallucination Risk
Simulating unconstrained generative arithmetic...
Detecting arithmetic discrepancies...
Locked JSON Fact Contract Sent to LLM:
Immutable Snapshot
// JSON contract