Security
Sensitive Data Inspect & Redact
Inspect or redact bounded text and JSON with deterministic rules, exact spans or JSON pointers, safe evidence, and no model call or payload storage.
CONTRACT
Predictable for agents. Legible for reviewers.
- REST
POST /v1/privacy/text/inspect- MCP tool
inspect_sensitive_text- Inputs
- Up to 256 KiB of supplied text or JSON, a selected subset of documented detectors, and an optional labeled, generic, or length-preserving replacement style.
- Returns
- CLEAR, REVIEW_REQUIRED, or BLOCKED with bounded findings, detector counts, exact text offsets or RFC 6901 JSON pointers, safe type-only previews, redacted output where requested, and an agent-readable next action.
- Latency
- Synchronous
- Usage unit
- Privacy operation
The focused suite also exposes /v1/privacy/text/redact, /v1/privacy/json/inspect, and /v1/privacy/json/redact plus the separate /mcp/privacy server. It detects a documented set of formatted emails, phones, U.S. SSNs, Luhn-valid payment cards, checksum-valid IBANs, IP addresses, URL credentials, common key/token formats, private-key blocks, and sensitive JSON field names. It is deterministic and non-exhaustive: it does not identify every person, postal address, health fact, biometric, free-form secret, or jurisdiction-specific personal-data category, and it does not certify compliance.
AGENT-READABLE EXCEPTIONS
PRIVACY_INPUT_TOO_LARGEPRIVACY_JSON_TOO_DEEPPRIVACY_JSON_NODE_LIMIT_EXCEEDEDPRIVACY_CANDIDATE_LIMIT_EXCEEDEDPRIVACY_FINDING_LIMIT_EXCEEDEDPRIVACY_OUTPUT_TOO_LARGESENSITIVE_DATA_FOUNDPOST /v1/privacy/json/redact
Authorization: Bearer apc_...
Content-Type: application/json
{
"value": {
"customer_email": "sam@example.com",
"authorization": "Bearer eyJhbGciOiJIUzI1NiJ9.payload.signature",
"note": "Keep this text"
},
"replacement_style": "LABELED"
}curl --request POST \
--url 'https://api-production-9502.up.railway.app/v1/privacy/json/redact' \
--header 'Authorization: Bearer ${AP_CONTROL_API_KEY}' \
--header 'Content-Type: application/json' \
--data '{
"value": {
"customer_email": "sam@example.com",
"authorization": "Bearer eyJhbGciOiJIUzI1NiJ9.payload.signature",
"note": "Keep this text"
},
"replacement_style": "LABELED"
}'{
"status": "COMPLETED",
"decision": "REVIEW_REQUIRED",
"finding_count": 2,
"counts_by_detector": {
"EMAIL": 1,
"SENSITIVE_FIELD": 1
},
"findings": [
{
"id": "EMAIL-001",
"detector": "EMAIL",
"category": "CONTACT",
"severity": "MEDIUM",
"confidence": "HIGH",
"rule_id": "EMAIL_ASCII_V1",
"start": 0,
"end": 15,
"json_pointer": "/customer_email",
"safe_preview": "[REDACTED:EMAIL]",
"evidence": {
"match_length": 15
},
"next_action": "Confirm the contact detail is authorized or use the redacted content."
},
{
"id": "SENSITIVE_FIELD-002",
"detector": "SENSITIVE_FIELD",
"category": "CREDENTIAL",
"severity": "HIGH",
"confidence": "HIGH",
"rule_id": "SENSITIVE_FIELD_NAME_V1",
"start": null,
"end": null,
"json_pointer": "/authorization",
"safe_preview": "[REDACTED:SENSITIVE_FIELD]",
"evidence": {
"field_name": "authorization",
"field_category": "credential",
"value_type": "str",
"match_length": null
},
"next_action": "Review the named field and use the redacted value unless explicit handling is authorized."
}
],
"issues": [
{
"code": "SENSITIVE_DATA_FOUND",
"severity": "HIGH",
"message": "The selected deterministic rules found 2 sensitive-data candidates.",
"evidence": {
"finding_count": 2,
"detectors": [
"EMAIL",
"SENSITIVE_FIELD"
]
},
"next_action": "Use the redaction operation or explicitly authorize each required value before continuing."
}
],
"ruleset_version": "2026-08-01",
"scan_complete": true,
"redacted_value": {
"customer_email": "[REDACTED:EMAIL_1]",
"authorization": "[REDACTED:SENSITIVE_FIELD_2]",
"note": "Keep this text"
},
"replacements_applied": 2,
"replacement_style": "LABELED",
"next_action": "Use only the redacted JSON and review every changed JSON pointer before consequential use."
}Use Sensitive Data Inspect & Redact through RapidAPI.
Choose a capped marketplace plan, copy the generated request, and route the structured result through your own authorized workflow.