What it does

dlt (data load tool)

You help build data ingestion pipelines using dlt. Use this when the user wants to extract data from APIs, databases, or files and load it into a warehouse or lakehouse.

Commands

Initialise a new pipeline

dlt init <source_name> <destination_name>

Example:

dlt init sql_database duckdb

Run a pipeline

python <pipeline_script>.py

Check pipeline status

dlt pipeline <pipeline_name> info

List loaded tables

dlt pipeline <pipeline_name> show

Pipeline Patterns

Basic API source

import dlt
@dlt.source
def my_api_source(api_key=dlt.secrets.value):
    @dlt.resource(write_disposition="replace")
    def customers():
        response = requests.get("https://api.example.com/customers",
                                headers={"Authorization": f"Bearer {api_key}"})
        yield response.json()
    return customers
pipeline = dlt.pipeline(
    pipeline_name="my_api",
    destination="duckdb",
    dataset_name="raw",
)
load_info = pipeline.run(my_api_source())
print(load_info)

Incremental loading

@dlt.resource(write_disposition="merge", primary_key="id")
def orders(updated_at=dlt.sources.incremental("updated_at")):
    params = {"since": updated_at.last_value}
    response = requests.get("https://api.example.com/orders", params=params)
    yield response.json()

SQL database source

from dlt.sources.sql_database import sql_database
source


## When to use it

When users need help creating, running, or managing Dlt pipelines and related tasks.


## What's included

- Scripts: none
- References: none


## Compatible agents

Likely compatible with general coding assistants and CLI-capable agents (Copilot, Codex, Gemini).