Product · SynxML
In-database ML, coming 2H 2026.
SynxML brings model training, retrieval, and vector search into the warehouse, so your features, embeddings, and predictions live next to the data that produced them. We're shipping SynxDB Cloud first; SynxML follows in the second half of the year.
What to expect
- Native vector search. Store and query embeddings alongside the features that generated them, no separate vector DB to keep in sync.
- SQL-first feature engineering. Build features as views, materialize them on demand, version them with the rest of your data model.
- RAG without the glue code. Retrieval for LLM applications, served from the warehouse your analytics already runs against.
- Training hooks to your model framework. We're not reinventing model training, we're making the warehouse a first-class participant in it.
Why not now
We're focused on SynxDB Cloud through the first half of the year. AI/ML needs the warehouse underneath it to be rock-solid first, otherwise the model is only as good as the features, and the features are only as good as the storage engine. Get that right, then expand.
Spin up a warehouse. Run a query. See for yourself.
Early access is open. A starter cluster is free while we're in preview, no credit card, no sales call.