Tonic Textual

The Tonic Textual Native App is a modern synthetic data platform for unstructured data, engineered for complex handling of free-text data. It integrates into data and ML pipelines and uses custom NER models to detect and redact sensitive data from 25+ entity types. Textual's synthetic data engine then re-seeds redactions with contextually relevant synthetic data, ensuring data privacy while maintaining data utility for model training, LLMOps, QA testing, and secure document sharing. Textual safeguards your data from breach, exfiltration, and model memorization and is an important asset for responsible AI development and data stewardship.<br/>

The Tonic Textual Native App features two main components: a detection service and a redaction service. When you run the TEXTUAL_REDACT function, text data in your Snowflake table is processed and a new privacy-preserving result-set is returned with redacted or synthesized text. Under the hood, our proprietary NER models are springing into action, identifying sensitive information which is then replaced with redacted or synthesized versions. It also utilizes Snowflake’s new GPU compute pools for exceptional performance at Snowflake scale. What makes this especially cool? Your data never leaves Snowflake, leveraging Snowflake’s robust security posture to maintain your data’s confidentiality and integrity.

<p><br/></p>

Setting up Tonic Textual is a breeze. A few commands to set up a compute pool and permissions, and that’s it—you're all set to start redacting text right within your Snowflake workflows. It’s as simple as clicking "Install" from the Snowflake UI, and the app is ready to go. You control who accesses the app and how it’s used, all with minimal setup fuss.