Safety monitoring
Unsafe-output classification runs even on the Free plan with storage-free scanning, and personal data is double-checked by SDK-side masking plus an LLM second pass.
Before / after PII redaction (example)
Before redaction (original prompt)
Inquiry from Taro Yamada ([email protected], +81-80-1234-5678). Card 1234-5678-9012-3456, address 1-2-3 Ginza, Chuo-ku, Tokyo.
After redaction (what gets stored)
Inquiry from Taro Yamada ([REDACTED_EMAIL], [REDACTED_PHONE]). Card [REDACTED_CC], address 1-2-3 Ginza, Chuo-ku, Tokyo.
Safety classification (example)
Scanned without storing plaintext — only the verdict labels are recorded.
Context-dependent PII that remains — like names and addresses — is caught by the LLM second-pass audit.
SDK & LLM double check
An LLM catches what the SDK regex missed via a cron-based second pass.
1SDK CHECK
Masked before sending
InstantRegex detection masks values before they leave.
2LLM CHECK
Second pass catches leftovers
Every 15 minRechecks with context and raises an alert on misses.
PII covered
A background sweep follows up so risks are not missed.
Consent + delete rights
Plaintext storage requires two-step consent. Deletion goes through a dedicated endpoint.
No plaintext stored
DEFAULTPlaintext is not stored by default.
Storing requires consent
Two-step consentEnabling storage requires a two-step consent.
Delete anytime
A dedicated endpoint deletes it for good.
DELETE /v1/records/:idUnpermitted decryption is structurally prevented.
Audit log
Every decryption is logged with who and when. Unauthorised decryption is structurally blocked.
Every decryption access is logged and protected