PrismoPrismo
Log in
Enterprise-Grade Security

Your data is protected
at every layer

Prismo is built with security-first architecture. We encrypt everything, log nothing by default, and give you complete control over your data.

SOC 2 Type II (in progress)
GDPR Compliant
256-bit Encryption

End-to-End Encryption

All data is encrypted in transit using TLS 1.3 and at rest with AES-256. Your API keys are protected with unique per-user encryption keys derived from secure key management.

  • TLS 1.3 for all API connections
  • AES-256-GCM encryption at rest
  • Per-user derived encryption keys

Zero-Logging Mode

Enable zero-logging for sensitive requests. We'll only store essential metadata (token counts, costs) — never your prompts, responses, or any content.

// Enable for any request
{
  "model": "gpt-4o",
  "messages": [...],
  "no_log": true
}
  • Per-request privacy control
  • Only metadata retained for billing

Data Handling

Prismo operates as a pass-through proxy. Your prompts flow directly to LLM providers and responses return directly to you. We never train on your data or share it with third parties.

  • No training on your data — ever
  • Semantic cache isolated per user
  • Configurable cache TTL (default 7 days)

Compliance

We're committed to meeting enterprise compliance requirements. Our infrastructure is designed to support regulated industries and data protection standards.

  • SOC 2 Type II (in progress)
  • GDPR compliant data handling
  • Data Processing Agreement (DPA) available

Complete Transparency

Every API response includes headers showing exactly what Prismo did. No black boxes — you always know what happened.

Response Headers
HTTP/1.1 200 OK

X-Prismo-Cache: hit
X-Prismo-Cache-Confidence: 0.9847
X-Prismo-Requested-Model: gpt-4o
X-Prismo-Actual-Model: gpt-4o-mini
X-Prismo-Routed: true
X-Prismo-Latency-Ms: 234
X-Prismo-Cache

Cache status: hit, miss, or skipped with reason

X-Prismo-Routed

Whether model was optimized for cost savings

Cache-Confidence

Semantic similarity score (0.0 - 1.0)

Latency-Ms

Total request processing time in milliseconds

Questions about security?

We're happy to discuss our security practices, provide documentation, or schedule a call with our security team.

Contact Security TeamView Documentation