Open-source Ruby tool built by Cybergeizer (led by Sergey Sergyenko) to obfuscate real production data rather than just seed fake data. The name is an English word for an animal that grazes ('eats a bit here and there'). In its 0.0.1 version Grazer: (1) scans existing models and generates per-model config files describing fields; (2) lets developers define rules per field — which are sensitive, and what obfuscation strategy to use (Faker generators, regional constraints, uniqueness preservation); (3) extracts the obfuscated data as a dump of SQL insert statements (full dataset or a sliced sample) so nothing sensitive leaves the server as a raw dump; (4) includes a consistency-tracking job that periodically diffs new/updated real records and updates only the delta in the obfuscated destination. Does not yet automatically propagate field dependencies (e.g. encrypted passwords referencing other fields) — that still has to be done manually. Used at Cybergeizer to make a HIPAA-compliant healthcare app usable for BI, analytics, staging on Heroku, and user-flow testing.