Tonic.ai product updates: April 2024

We're excited to share the latest updates and announcements designed to improve your experience with our products. This month's issue includes:

• Support for SQL Server on Tonic Ephemeral ✅
• File connector enhancements on Tonic Structural 📂
• Support for Db2 LUW on Tonic Structural 🙌
• LLM synthesis on Tonic Textual 💫
• and expanded access to LLMs on Tonic Validate 🎡

Support for SQL Server on Tonic Ephemeral ✅

Our new solution for spinning up (and down) ephemeral databases on demand now supports SQL Server, in addition to PostgreSQL and MySQL. Tonic Ephemeral makes it easy for developers to spin up fully populated test and development databases for ephemeral test environments so you can work more efficiently while keeping costs under control. You can set Ephemeral as your destination database within Tonic Structural to generate and output de-identified SQL Server data directly into an ephemeral database. This powerful pairing is particularly effective for subsetting use cases, when your developers need rapid access to a small chunk of data for a small amount of time. We're currently offering free trials of Ephemeral—sign up here or book a demo to learn more.

‍

File connector bonanza (including Parquet support!)

We released a whole slew of improvements to our file connector recently, all in the name of efficiency:

• Selecting a directory or prefix: You can now select a directory or prefix instead of selecting individual files. Save time by targeting groups of files instead of individual files. This enhancement also offers you the option to only process new files added to the directory since the previous generation, enabling you to automate data generation for your directories that are regularly receiving new files.
• Download files from job details view: When generating for local file workspaces, you can now download the de-identified files from the job details view. No more searching across the UI to find the output.
• Parquet support! Last but certainly not least, we added support for Parquet formatted files through the file connector. Upload and de-identify your Parquet files directly, no middle-man data store needed.

Keep File Connector in mind for additional teams in your organization, or to treat file pipelines with the same care for safe, realistic test data as your databases.

‍

Support for Db2 LUW on Tonic Structural 🙌

Tonic Structural now offers native support for Db2 LUW, the relational database management platform from IBM. Connect to your Db2 LUW databases to de-identify and synthesize your data for secure use in your lower environments without sacrificing data utility along the way. And if you’re working with another data source, as well, you can ensure consistency across database types by configuring the TONIC_STATISTICS_SEED in your Structural environment settings. Subsetting for Db2 LUW is coming soon as a fast follow. 👀

Using another flavor of Db2 that you’d like to connect to Tonic Structural? Book time with our team to tell us more about your needs in this space.

‍

LLM synthesis on Tonic Textual 💫

Tonic Textual, our solution for standardizing, enriching, and protecting your unstructured data for AI development, now supports advanced LLM synthesis as an option for realistic data de-identification. Instead of simply redacting your sensitive data, leverage Textual’s proprietary synthetic data capabilities to replace original values with contextually relevant and accurate synthetic values. Textual’s LLM synthesis maintains the semantic realism of your original data, preserving its shape, utility, and meaning—while also preserving data privacy.

Additionally, Textual also supports image file types like jpg, png, and tif, enabling you to seamlessly extract text from images to feed your ML pipelines.

Working on RAG or LLM development and in need of a solution to make your unstructured data AI-ready? We’d love to speak with you. Connect with the Textual team here.

‍

Expanded LLM support on Tonic Validate 🎡

Our RAG evaluation platform, Tonic Validate, has expanded its support to allow you to use a variety of LLMs for evaluation. Previously, we only supported models from OpenAI, but now users can select from among the leading LLM providers, including Google (Gemini), Azure OpenAI, Anthropic (Claude), Ollama, Together AI, Mistral, AWS Bedrock, and Cohere. Having the flexibility to choose the model that works best for your business enables you to better manage cost and performance based on the task at hand. Whether you’re building a customer-facing or internal RAG system, Validate gives you the tools you need to create benchmarks and monitor how your RAG system outputs are changing over time with real-world usage. Sign up for a free trial and let us know what you think.

‍

Small Updates; Big Impacts

Often it's the little things that matter most. Here's a round up of our smaller releases:

• Output to repos 🤝 SQL Server: You can now write destination data from SQL Server to a container repository, for more flexible data provisioning.
• Just when you thought our Regex mask generator couldn’t get any cooler, we’ve added the UUID key to its list of subgenerators options. 😎
• Check out Structural’s new API endpoint to resolve all schema changes in a workspace. Choose whether to resolve only conflicting changes, only notifications, or all of the schema changes, for more efficient automation.
• Subsetting on Snowflake now respects virtual foreign keys, offering you more control in crafting your Snowflake subsets. ❄️

As always, we'd love to hear your feedback on our products. What do you need? What do you love? What could be better? Send us a note at hello@tonic.ai! And for all the latest updates, be sure to check out our complete release notes in our product docs.