Velvet is a technical product that maps to your unique data model. Learn about Velvet’s AI SQL editor and how your whole team can use it to build better features, faster.
It’s a tedious and time-consuming process to set up data infrastructure. We manage the complexity on our end so it’s simple for you to unify volumes of data from your internal databases, third-party sources, and product webhooks. Depending on your specific infrastructure, you should be able to get set up in a few hours.
Everyone on your team will have access to the sources you connect, reducing your overhead so you can focus on higher-priority product development work.
We use ClickHouse's Table Engines to enable read-only access to your database. Choose which tables you want to give access to, or create read-only database users. Each table you add maps to a table in Velvet.
We’ve built API-driven syncing services so you don’t have to. Just drop in an API key and start syncing. Velvet will capture every event you specifiy in a queue, without slowing down your application database.
A good starting point would be something like this:
Track every single product event with a collection queue. This feature is super flexible — send us any JSON and we’ll write it to your events table for that source. Events can be captured from webhooks, or using the Velvet events API.
Then write complex queries across data sources like - “Show me super users and an array of video recommendations for each”.
You'll get a database per workspace, and a table for each source.
We’ve built an AI interface so you can use natural language to write complex SQL across disparate and unstructured data sources. Our system introspects on your data to write SQL, heals to resolve errors, and returns results.
Under the hood, we use ClickHouse to bring all your data together into one queryable interface. Each organization's workspace a ClickHouse database, and each source is a table.
When you ask your data a question, we return results in the form of data tables or numerical analytics. Use results to understand your system or develop new features.
There are two types of results you can request:
Write queries that return lists or analytics results.
Define, track, and ship features in one place - and spend less time orchestrating data infrastructure. With Velvet, your whole team has autonomy to experiment with data.
Use Velvet to reduce costs and build faster: