Velvet is a database service with an AI-powered SQL editor on top. We design Velvet for engineering and product teams at high-growth startups.
We map to the complexity of your data model and surface SQL directly, so you’ll want an understanding of your system’s architecture to use the AI editor. The core product is designed for engineers and technical product managers who are building and maintaining features.
You can invite non-technical marketing and CX stakeholders as read-only users. They’ll be able to view and use already defined queries.
This means you can send us volumes of data that can be queried in milliseconds. Each workspace is a database and each source is a table in our system. You can create as many workspaces and tables as you need.
We leverage AI models as an accessible interface for data management. Velvet introspects on the data you connect to write complex SQL. We currently use OpenAI, which is SOC 2 Type II compliant.
You can either prompt the AI editor with natural language, or edit manually. Run the editor to return results from your live data, and keep iterating until you’re happy with the data set.
We can ingest data from any source. We recommend starting with your application database, payment platform, and product analytics. See our list of supported data sources here. Email us with any additional data source you want supported, and we’ll create a guide.
Read our docs to get started. We can also support you with data engineering resources so you can focus on building your core product.