Basics
Introduction
PromptQL is powered by Hasura DDN, which enables a powerful, yet simple-to-create AI-enabled API on your data.
Each tutorial in this section will guide you through creating your PromptQL project on top of your preferred data source. First, let's go over the broad stokes of how it works.
You have data
Your data can live in a variety of places - relational or non-relational databases, OLAP or OLTP databases, vector stores, third-party services, or output by your code. With PromptQL, you can talk to all of it from a single endpoint.
You connect it to PromptQL
Regardless of where your data lives, you'll connect it to your API using a native data connector. These connectors will introspect your data sources and generate metadata that Hasura DDN will use to build and power your PromptQL application.
You configure your API in metadata
This semantic metadata layer is written as Hasura Metadata Language (HML), which is an extension of YAML. You can use the DDN CLI and your text editor to fine-tune and configure the API to your liking; this includes easily creating metadata objects for permissions, relationships, configuring authentication, etc.
You talk to your data
To convert your metadata into something that you can talk with, you'll create a build - an immutable snapshot of your API at that point in time which the Hasura Engine can run and via PromptQL. You can then use the PiQL (pronounced pickle) - our GUI - to talk to your data, or integrate it with your client apps via the PromptQL API.
Get started with your favorite source
See the whole list in the sidenav to the left!