What synthetic mode does
Redaction removes context. Replacement tokens keep consistency but still look fake. Synthetic mode is for demos, testing, training samples, and LLM prompts where the document needs to read naturally without revealing the real person, company, account, or address.
How Piixie uses Faker
Piixie asks the local model to identify entities and return a plan: entity type, locale, generator, original text parts, and transforms. Piixie then runs that plan locally through a bundled @faker-js/faker runtime hosted inside a pure-Go JavaScript VM.
The fake values are produced on your machine. Piixie does not call a cloud data-generation API.
Supported generator families
People and organizations
Names, first names, last names, titles, possessive forms, company names, and email-friendly name parts.
Contact and location data
Street addresses, cities, states, countries, postal codes, phone numbers, URLs, IP addresses, and email addresses.
Identifiers and finance
Account numbers, IBANs, card numbers, UUIDs, alphanumeric IDs, words, and birthdates.
Consistency rules
Every logical entity gets a cached Faker profile. If a document contains a person's full name in one paragraph and the same person's last name later, Piixie reuses the same fake profile so the replacement stays consistent.
- Multi-line addresses are generated from one address profile.
- Emails can reuse the synthetic person profile for realistic local parts.
- Company-related domains can use company-shaped names where possible.
- Locale hints are normalized to English or Spanish Faker data today.
Shape preservation
Some values are useful because their shape is meaningful. Piixie can preserve separators, casing, digit counts, and broad format patterns while still replacing the real value.
14/03/1982 stays in day/month/year style, an ID with letters and digits keeps the same rough pattern, and a work email can keep the same style without keeping the real name.When to use it
- Preparing safe examples for LLM prompts.
- Creating demos that look real without production PII.
- Building test fixtures from real document shapes.
- Sharing support cases where redaction would remove too much context.