Gamut
Get in Touch
ACRA  ·  Companies House  ·  SEC EDGAR

The auditable record
behind every entity decision.

Every claim traced to its primary source. Every score you can take apart and audit.

The verification layer that runs on top of your data stack — making entity assertions defensible for regulated decisions.

Get in Touch See How It Works ↓

Claim → Extraction → Source Match → Assertion Verdict

01

Discover

AI-powered entity discovery across jurisdictions. Multi-source search with intelligent deduplication and mandate-fit scoring.

02

Verify

Government registry verification plus cross-source claim validation. Every claim tagged by source tier — registry, press, web.

03

Decompose

Deterministic confidence scoring. Users see the evidence, not just the conclusion. Auditable by design.

Three Jurisdictions. Live.

🇸🇬
ACRA

Singapore

Every registered company in Singapore — live registry verification via ACRA.

Live
🇬🇧
Companies House

United Kingdom

Every active company in the United Kingdom — Companies House REST API.

Live
🇺🇸
SEC EDGAR

United States

Every SEC filer in the United States — EDGAR integration, zero-cost, no auth required.

Live

One Engine. Three Verticals.

Private Company Diligence
High-conviction diligence where primary registries are hardest to reach.
Deep private-company verification across US, UK, and Singapore jurisdictions
Every claim connects to a live primary registry at query time — not a precomputed graph
Built for cross-border deals where public data runs out
Compliance & KYC
Third-party risk and onboarding decisions backed by primary-source evidence.
Adverse signals — litigation, sanctions, regulatory action — isolated by primary-source trust tier
Scoring template configures to your compliance policy — not a fixed vendor risk ontology
Full audit trail for every onboarding decision
Insurance Underwriting & Governance
Underwriting signals, traced to source — and the governance layer underneath.
Litigation, regulatory action, financial distress — each isolated and tiered by primary-source trust
Configures to your risk model — not a fixed vendor ontology
Full audit trail for every underwriting decision

Each vertical is a configuration — not a separate product. New verticals ship without pipeline changes.

How Gamut Works

PitchBook
D&B / Dun & Bradstreet
CRM / MDM
Internal Data
Gamut
Registry Match Claim Validation Confidence Scoring
Investment Committee
Compliance Sign-off
Policy Binding
Counterparty Onboarding

Gamut is not a destination portal or a closed data graph. It is a single, auditable record that sits between your data stack and your applications. Whether verifying a private company for a cross-border venture investment or tiering litigation risk for an underwriting decision, the architecture is identical: source-attributed claims, reproducible scoring, and a verdict you can trace end to end. Same engine. Three verticals.

Confidence You Can Decompose

Registry Match 30 pts
Entity confirmed in a government registry
Source Diversity 30 pts
Claims corroborated across multiple source tiers
Cross-Source Consistency 25 pts
No contradictions between sources
Entity Completeness 15 pts
Key fields present and non-empty

No LLM in the scoring path. Pure deterministic Python. Every point traceable to an observable fact.

score = registry×0.30 + diversity×0.30 + consistency×0.25 + completeness×0.15

Intelligence Briefings

Verified startup landscapes, auto-generated from pipeline output.
Every claim traceable to its source.

Subscribe

RSS Feed Apple Podcasts & Spotify — coming soon
Open Source

Verity — the authenticity
layer, open source.

The three-layer content authenticity pipeline that powers Gamut's signal filtering is available as a standalone open-source tool. Monitor any topic, filter AI-generated noise, and surface only what's credible — with your own LLM and your own search provider.

gamutagent/verity Apache 2.0
Layer 1 — Source Reputation
YAML-driven domain trust tiers. Reuters scores differently than a PR wire. Zero API cost.
Layer 2 — Content Heuristics
7 deterministic Python checks: AI hedge phrases, uniform sentence length, missing bylines, clickbait patterns. Zero API cost.
Layer 3 — LLM Detection (optional)
One API call per article. Uses your existing scoring key. Disable for high-volume runs.
composite = source×0.45 + heuristic×0.55
# weights configurable in config.yaml
Get in Touch

See the Evidence
Yourself.

Request access to run a live verification. No demo theater — real queries, real registries, real confidence scores.

Get in Touch Read the Architecture →