Forge

The Problem

AI + Financial Modeling = Disaster

LLMs hallucinate numbers (confident but wrong)
Excel files cost tokens (MCP tools add ~12k overhead)
CSV loses formulas (flat data only, no logic preserved)
JSON bloats token counts (40% more than needed)
FP&A platforms cost $50K-$200K/year

What if financial models were text files that AI could read,
with formulas that actually calculate?

The Solution

# budget.yaml - Your financial model as code
_forge_version: "1.0.0"

assumptions:
  revenue:
    value: 1000000
    formula: null
  cogs:
    value: 400000
    formula: null
  gross_profit:
    value: null
    formula: "=revenue - cogs"
  gross_margin:
    value: null
    formula: "=gross_profit / revenue"

Write in YAML. Evaluate with Forge. Export to Excel.
No cloud. No subscription. No lock-in.

Capabilities

📊

173 Functions

Excel functions + Monte Carlo distributions + FP&A (NPV, IRR, XNPV, XIRR, VARIANCE, BREAKEVEN).

🎲

Monte Carlo Simulation

Normal, Triangular, Lognormal, PERT, Discrete distributions. Latin Hypercube sampling.

🌳

Decision Trees

Sequential decisions with backward induction. Optimal path and expected value calculation.

💎

Real Options

Value flexibility to defer, expand, or abandon. Black-Scholes and Binomial Tree pricing.

🌪️

Tornado Diagrams

Sensitivity analysis showing which inputs drive output variance. One-at-a-time analysis.

🔁

Bootstrap Resampling

Non-parametric confidence intervals from historical data. Percentile and BCa methods.

🕸️

Bayesian Networks

Causal dependency modeling with belief updates. Evidence propagation for risk analysis.

🔄

Excel Import/Export

Import .xlsx to YAML. Export YAML to .xlsx with working formulas. Bidirectional.

🤖

AI-Native Format

LLMs trained on 13M+ YAML files. Formulas preserved. Feed to Claude/GPT as plain text.

📝

Git-Trackable

Version control for financial models. Diff, branch, PR, blame. Real audit trail.

⚡

Local-First

Runs on your machine. No cloud. No telemetry. Works air-gapped.

🔬

External Validation

Every formula validated against Gnumeric (academically proven accuracy). Statistical functions validated against R (the gold standard). 1,700+ unit tests. Zero tolerance.

Capability	Excel	FP&A Platforms	Forge
Git version control	Binary files	Proprietary	Native YAML
AI token efficiency	~12k MCP overhead	Cloud API	Native text format
Works offline	Yes	Cloud required	100% local
Formula evaluation	Yes	Yes	173 functions
Annual cost	$150+	$50K-$200K	R&D (inquire)

Try It

Try the demo (47 functions, 70 tests). Enterprise: 173 functions, 2,703 tests with 100% test integrity.

# Clone and run the demo
git clone https://github.com/royalbit/forge-demo
cd forge-demo
./run-demo.sh

No installation required. The script downloads the binary and runs validation tests.

View on GitHub