Every API call through ACAI generates compliance evidence automatically. Below are real report formats — the same outputs your team downloads after signing up.
Period: 2026-01-01 — 2026-03-31 · Generated: 4/1/2026
All API keys scoped to individual users. MFA enforced via Entra ID. No shared credentials detected.
SHA-256 checksums on all audit log entries. Tamper detection enabled. Zero integrity violations.
TLS 1.3 enforced on all endpoints. HSTS enabled. No plaintext transmissions detected.
90-day audit log retention. 142,847 events captured. Export available in JSON/CSV.
Bearer token auth with per-key rate limits. Entra ID MFA for dashboard. Token rotation every 90 days enforced by policy.
PII detection enabled (14 patterns). 3 requests flagged with unredacted SSN — auto-redacted before model. Recommend client-side input validation.
Zero breaches. Incident response runbook verified and tested (last: 2026-02-15). Notification timeline: 60 days.
BAA executed with Microsoft (Azure AI Foundry). Sub-processor list documented. ACAI BAA available for customer execution.
Every API request is logged with model, token count, PII detection result, data classification, and compliance verdict.
| Time | Model | Tokens | PII | Classification | Verdict |
|---|---|---|---|---|---|
| 14:31:42Z | gpt-4.1-mini | 847 | none | internal | PASS |
| 14:31:38Z | claude-sonnet-4-5 | 1,203 | SSN (redacted) | confidential | PASS |
| 14:31:35Z | deepseek-r1 | 2,104 | none | public | PASS |
| 14:31:29Z | gpt-4.1-mini | 512 | email (redacted) | confidential | PASS |
| 14:31:22Z | phi-4 | 384 | none | internal | PASS |
One ZIP file your auditor can review offline — generated automatically from your API usage.
Framework-specific control mapping with pass/warn/fail verdicts and remediation guidance.
Every API request: model, tokens, PII status, data classification, timestamp, correlation ID.
Aggregate stats: patterns detected, redaction actions taken, false positive rate.
TLS version, cipher suites, key rotation schedule, certificate chain.
Token consumption by model, cost breakdown, rate limit utilization, error budgets.
Active guardrails, content safety settings, data classification rules, retention policy.
Sign up free, make API calls, and download real evidence packs in minutes.