Beyond Copilots · On-demand

Watch dltHub, MotherDuck and Lightdash take a live API from ingestion to a governed AI answer, using agents at every layer and one terminal throughout.

Recorded live on 18 August 2026

Watch the full build

See the stack come together from the first API request to the final governed answer.

Want to see what this could look like on your data?

Start for free

The query was never the hard part.

The hard part is everything around it: getting data into the warehouse, understanding the schema, defining metrics and giving an agent enough context to return an answer you can trust. This session showed what happens when agents can work across the whole stack instead of stopping at text-to-SQL.

A terminal running the live build: dlt init api, a motherduck query and lightdash metrics, each step ending in a governed answer

Build it yourself

  1. 01Get the source codeSee the pipeline, configuration and files used during the live build.
  2. 02Build the pipeline with dltHubGenerate and deploy an ingestion pipeline from an API.Read the dltHub guide
  3. 03Run it on MotherDuckLoad and query the data in a serverless DuckDB warehouse.Read the MotherDuck guide
  4. 04Add governed analytics with LightdashDefine metrics in code and connect your AI agents to a governed semantic layer.Read the Lightdash guide

Built live by

  • Alena Astrakhantseva
    Alena AstrakhantsevaSenior Solutions Engineer
  • Jacob Matson
    Jacob MatsonDeveloper Relations
  • Oliver Laslett
    Oliver LaslettCo-founder

Give your agents access to your metrics in minutes.

SCORE 0