Cyphr chosen to power $5.9B in economic mobility capital.

A model built for the 80% bureau models can't see.

Cyphr Core is a domain-specific model trained on small-business capital signals — banking, licensing, operations, and program outcomes. The same readiness layer powers free business scores, lender underwriting, and ecosystem dashboards.

SOC 2 certified1071 readyModel governanceBias monitoring
cyphr core · readiness outputLive
{
  "model": "cyphr-core",
  "score": 244,
  "dimensions": 9,
  "signals": ["banking", "licensing", "operations"],
  "explainable": true,
  "compliance": ["1071", "SSBCI"],
  "consumer_credit_used": false
}
Three principles

Built differently on purpose.

Cyphr Core wasn't adapted from a consumer credit model. It was built for small-business capital from the first signal forward.

01

Cash-flow native

Built from the ground up on small-business banking signals — not consumer credit. The 80% of Main Street that's invisible to bureau models is the population we were designed for.

02

Outcome-trained

Calibrated against real funded loans, real defaults, and real program outcomes across four states. The model learns from what actually happened, not what a credit file predicted.

03

Auditable by design

Every score decomposes into the nine readiness dimensions that produced it. Every decision carries the evidence a regulator, an underwriter, or a business owner can read.

The transformation

Signals in. Readiness out.

Banking rhythm, government records, operational evidence, and program outcomes flow into Cyphr Core — the same readiness layer that powers a free business score, a lender's underwriting decision, and a state agency's compliance report.

Inputs
Banking signals12.4k/s
Government records3.1k/s
Operational evidence8.7k/s
Program outcomes442/s
Readiness engine
Outputs
Readiness score244
Fundable range$120–280K
Packet completenessReady
1071 / SSBCI fields✓
What we don't do

The model is defined as much by what it refuses to use.

—

No consumer credit data. Ever.

Bureau pulls don't fit Main Street. We score the business, not the owner's personal credit history.

—

No scraped social or behavioral signals.

What you posted, what you searched, who you follow — none of it belongs in a credit decision.

—

No black-box decisioning — every score is explainable.

If a score can't be defended in plain language, it can't be defended to a regulator either.

—

No selling data. The business owns their profile.

The readiness profile belongs to the business. They decide who sees it and when.

Governance

Built to be audited.

Every score the model produces is explainable, every decision is logged, and every cohort is monitored for drift and disparate impact.

SOC 2

Certified

SOC 2 certification complete with a top-tier firm. Controls cover data handling, access, and model lifecycle.

1071

Native compliance

Required fields populated at intake. Reporting becomes an output of the system, not a quarterly scramble.

FAIR

Bias monitoring

Continuous disparate-impact testing across demographic and geographic cohorts, with documented remediation.

One model. Three perspectives on Main Street credit.

From free business scores to lender underwriting to ecosystem dashboards — all powered by Cyphr Core.