For engineers

Use a collaborative AI SQL editor

Build iterative queries on top of your database and unified data. Collaborate on real-time data features and make analytics accessible to the entire team.

Why Velvet

When tasked with building or analyzing a new feature - you spend valuable time piecing together data sources, managing in-memory "joins", and general maintenance.

Instead, use Velvet to query data in one unified workspace. Now any product stakeholder can write complex SQL using natural language. Return live data in milliseconds, experiment with new features, and build on top of our API.

Velvet makes everyone on your team a data engineer.

How it works

We've engineered a data pipeline so you don't have to.

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.

So you can write queries like this, without breaking a sweat.

1WITH 
2    toDateTime(date_sub(week, 1, now())) AS start_of_last_week,
3    toDateTime(now()) AS end_of_last_week,
4    (
5        SELECT groupArray(p.name) AS recommended_product_names
6        FROM postgres_products p
7        WHERE p.active = 1
8    ) AS recommended_products
9SELECT DISTINCT 
10    u.email, 
11    u.full_name, 
12    u.id,
13    recommended_products.recommended_product_names
14FROM postgres_users u
15JOIN postgres_subscriptions s ON u.id = s.user_id
16WHERE s.status = 'active'
17  AND s.current_period_start >= start_of_last_week
18  AND s.current_period_start <= end_of_last_week
19ORDER BY s.current_period_start DESC;

Don't know all the specifics of a data source's schema? No problem. Ask any natural language question about your product and our AI editor introspects on your data to write the SQL for you. We've built in "healing" on errors, and the editor recommends additional queries that may be of interest based on the shape and patterns of your data.

AI recommended queries illustration
Data sources

We recommend you start by connecting your database, then layer on additional data sources as you need them. Each source is a table that can be queried.

Velvet infrastructure

Velvet security

A guide to getting started for engineers

Connect your data sources

Our system can support most databases, data integrations, and events. If you don't see one of your required sources listed here, fill out this form (or email team@usevelvet.com) and we'll add support.

Databases
adding a new database table illustration

Read our docs to connect your database. It's a quick set up.

Fill out this form to request support for a database - we'll add it!

Third-party data sources
adding a new 3rd-party integration illustration

Read our docs to add third-party integrations. They're quick to set up.

Fill out this form to request support for a third-party tool - we'll add it!

Capture webhooks from anywhere
adding a new event illustration

Read our docs to capture events from any source. They're quick to set up.

Get started

We're here for you! Schedule a call to get a demo or for onboarding support.

Email us any time at team@usevelvet.com, or schedule a call here.

ai-first data platform

Make everyone on your team a data engineer.

Try Velvet for free

Articles to learn more

Product
Enter the Velvet challenge

Collect the most points for the chance to win a cash prize.

Engineering
Sync your database - How it works

Securely connect and query your database with Velvet.

Product
How we used Velvet during our product launch

Using Velvet to analyze our launch and automate onboarding.