Show HN: Fabi 2.0 – An AI analyst that connects to all your data sources
Today we’re launching Fabi 2.0
For the past year we’ve been working with a number of product, eng, marketing and data teams, and the common issue we’ve been seeing with all these teams, especially small and growing teams, is that getting basic answers from their data requires jumping between a bunch of different tools/copy-pasting data and spreadsheets, or building out a full BI stack.
With our latest release, users can not only connect Fabi to databases and data warehouses, but also to applications like HubSpot, Stripe, Shopify etc. We offer hundreds of connectors, and the AI makes it dead simple to start immediately building out analyses or dashboards.
A bit about what’s going on under the hood:
- We use Fivetran to connects to data sources (we tested at a number of different ETL solutions and Fivetran really stood out) - We pipe the data into a MotherDuck instance for each customer - We index data and metadata and store in our vector DB to get it ready for the AI. A lot of connectors just work without any further modeling, but some do require a bit of extra modeling. For those we automatically spin up dbt models to do some data wrangling under the hood - We offer an extra AI config layer for users to provide any additional context about their business and data source which they can write in plain English, which we’ve found to be really important to make it accessible to a wide set of teams
Once the data is ready, you can start talking to it to analyze it like you could before in Fabi with any traditional SQL DB.
You can give it a try here: [https://app.fabi.ai/](https://app.fabi.ai/home)