LLM Output Validation API

Your LLM ships output.
We make sure it's safe.

OutputGuard validates LLM responses in real time — catching PII leaks, prompt injection, policy violations, and schema errors before they reach your users.

Request API access → See how it works nFADP · FINMA-aligned · Swiss-hosted
POST /v1/validate
# Validate any LLM output in one call
curl https://outputguard.dev/v1/validate \
  -H "X-API-Key: og_your_key" \
  -d '{"text": "Send funds to IBAN CH56 0480 5000 0012 3456 7", "policy": "finance"}'

# Response
{
  "action": "blocked",
  "risk_score": 0.92,
  "issues": ["financial_pii", "iban_detected"],
  "audit_id": "evt_7f3a..."
}

What it checks

PII

Data leak detection

Names, emails, IBANs, AHV numbers, credit cards — caught before output leaves your system.

Injection

Prompt injection blocking

Detects injection attempts embedded in LLM outputs before they propagate downstream.

Policy

Finance & fiduciary rules

FINMA-aligned checks for investment advice, regulatory phrases, and fiduciary language.

Audit

nFADP audit export

One-click PDF/JSON audit reports for Swiss data protection compliance. SHA-256 hashed, no raw PII stored.

Review

Human review queue

High-risk outputs flagged automatically for human review. Full resolution tracking.

Schema

Output schema validation

Enforce structured output contracts. Catch malformed JSON and missing fields instantly.

Compliance coverage

nFADP / Swiss DPA FINMA guidelines NIS2 GDPR Swissmedic ISO 27001

Pricing

Free

€0

500 req / month

Get started

Developer

€19

10K req / month

Get started

Pro

€49

50K req / month

Get started

Scale

€149

250K req / month

Get started

Ready to validate your LLM outputs?

Get API access or request a pilot for your team. Swiss enterprise pricing available.

Request access →