Redacting PII

You can easily remove personally identifiable information (PII) or any other sensitive data when ingesting data into Faros AI.

PII scanning and stripping consists of the following settings:

• redact_fields_map: a JSON map where the key is a model name from the Faros Schema and the value is a list of the fields that one wants to be redacted. E.g. { “tms_Task”: [“description”] }
• redact_custom_replace (optional): the value to be used as replacement for any detected PII patterns. E.g. “REDACTED”
• redact_custom_regex (optional): extend the default built-in patterns with a list of additional custom regular expressions
• exclude_fields_map (optional): completely exclude the whole field from a record. E.g. { “tms_Task”: [“description”] }

These settings are available on both Sources and Webhooks. In both cases, records will be redacted before data gets ingested into Faros AI.

Sources

• Web application UI: available for every Source under Advanced Settings:

• Airbyte CLI / UI : via parameters:

Webhooks

When creating a webhook the request body can receive the stripper configuration as part of the settings field:

"settings": {
    "redactFieldsMap": {
        "tms_Task": ["description"]
    },
    "redactCustomReplace": "REDACTED",
    "redactCustomRegex": ["(\[a-z0-9_\\-.+]+)@\\w+(\\.\\w+)\*"]
 }