New · vario API

The vario engine, also as an API for forensic communications audits

Integrate risk analysis of anticompetitive conduct and corruption directly into your code. The same engine vario uses, available as a REST API with a single endpoint.

Create a developer account Talk to the team

250 free tokens · No credit card required

View documentation →

Full technical reference: endpoints, authentication, examples and rate limits.

Request

POST /api/v1/analyze
Authorization: Bearer vario_xxxxxxxxxxxxxxxxxxxxxxxx

{
  "text": "Tengo los precios de siempre.
          Habló con ellos, quedaron conformes.",
  "conduct_type": "price_fixing",
  "mode": "full"
}

Response200 OK

{
  "combined_score": 8.4,
  "should_alert": true,
  "severity": "high",
  "reasoning": "El texto evidencia coordinación…",
  "key_phrases": ["precios de siempre", "conformes"],
  "regulatory_frameworks": ["CADE", "COFECE", "FNE"],
  "recommended_action": "legal_review",
  "confidence": "high"
}

How it works

You send a text. The AI engine analyzes it and returns a structured risk signal in seconds.

Send the text to analyze

A POST to /api/v1/analyze with the communication text and your API key. Plain text, JSON, up to 8,000 characters.

Proprietary AI engine

The same AI engine that runs inside vario: filters noise, analyzes context and produces a structured risk judgment in seconds.

Structured risk signal

You receive a JSON with a 0–10 score, severity, conduct type, reasoning, key phrases from the text, and applicable regulatory frameworks.

Complete output in every analysis

Each response includes: risk score (0–10), severity (critical/high/medium/low), detected conduct type, model reasoning, key phrases from the text as evidence, applicable regulatory frameworks (CADE, COFECE, FNE…) and a recommended action.

combined_score

0–10

should_alert

boolean

severity

critical / high / medium / low

reasoning

string

key_phrases

string[]

regulatory_frameworks

string[]

recommended_action

legal_review / monitoring / no_action

confidence

high / medium / low

Use cases

Forensic investigations

Forensic firms processing large volumes of communications for regulatory investigations. Automated triage before human review.

E-discovery and litigation support

Integrate risk scoring into e-discovery workflows to prioritize relevant documents before legal review.

Legal review triage

Legal teams that need to prioritize human review across hundreds of thousands of messages exported from an Exchange server or Gmail.

Integration into your product

Compliance, GRC or audit platforms that want to add communications risk analysis as a native capability.

The text you send is never stored

The API is stateless by design. Each call is independent: the text you send is analyzed and discarded — it is never persisted in vario's databases. Designed so firms and legal teams can analyze privileged communications without transferring custody.

Ver arquitectura de seguridad →

Three modes for your use case

Screening mode

~50ms · 1 token/call · L1 layer only (keyword + embedding, no LLM). For mass corpus sweep and fast discard of low-signal documents.

Fast mode

~200ms · 2 tokens/call · L1 + L2 (light LLM). For moderate-precision classification after the initial screening pass.

Full mode

2–8s · 3 tokens/call · L1 + L2 + L3 (deep LLM). For high-precision analysis with full reasoning and key phrases as evidence.

Eleven conduct playbooks

Competition (price-fixing, bid-rigging, information exchange, group boycott), anti-corruption and integrity (bribery, conflict of interest, internal fraud), market integrity (market abuse, insider trading), and workplace integrity (harassment, whistleblower retaliation). Calibrated for LATAM regulators: CADE, COFECE, CNDC, FNE.

price-fixingbid-riggingbriberymarket-manipulationinformation-exchangegroup-boycott

Spanish, English and Portuguese, by design

The only public inference engine for antitrust/compliance conduct calibrated natively in Latin American Spanish, English and Brazilian Portuguese.

How do you access the API?

Two paths depending on how you use vario.

Pay-per-use pricing

Pay only for what you analyze. Screening 1 token · Fast 2 tokens · Full 3 tokens — choose the depth you need.

$0.15per token

250 free tokens when you create your account — one-time, included on all plans.

Create a developer account

The API output is a risk signal, not a legal determination. It does not replace legal counsel review and does not constitute evidence per se before regulators.

Start protecting your communications

Tell us about your organization and a member of the vario team will show you how it works.

Talk to the team

No commitment · The vario team will get back to you soon