Quickstart

Build and deploy your first Buster agent in just 10 minutes. You’ll need:

A GitHub account with admin access to a repository
A dbt project
A data warehouse (Snowflake, BigQuery, Redshift, etc)
Warehouse credentials

1. Connect your stack

Go to our website and create an account. When prompted, install the Buster GitHub App on your repository. Choose “All repositories”, or select specific repos.

You will then be prompted to connect your data warehouse. See the Data Warehouse integrations section for connection details specific to your warehouse (Snowflake, BigQuery, Redshift, etc.).

2. Create an agent

Create a file at .github/buster/agents/docs-updater.yaml in your repository:

docs-updater.yaml

name: docs-updater
description: Automatically updates dbt model documentation when models change

triggers:
  - type: pull_request
    events: [opened, synchronize]
    on_changed_files: "models/**/*.sql"

tools:
  preset: standard

prompt: |
  When dbt models change in this PR:
  
  1. For each changed model, profile it using the warehouse:
     - Get row count, column names, and data types
     - Calculate null percentages for each column
  
  2. Update the corresponding YAML file with:
     - Model description including purpose and approximate row count
     - Column descriptions with data type and null rate if > 5%
     - Use sentence case and end descriptions with periods
  
  3. Run `dbt parse` to validate the YAML syntax
  
  4. If validation passes:
     - Commit the updated documentation to this PR branch
     - Add a comment: "📝 Updated documentation for [model names]"
  
  5. If validation fails:
     - Do not commit changes
     - Comment on the PR with the validation error and suggested fix

Commit and push to your main branch. Buster will automatically detect and deploy your agent within seconds.

3. Test your agent

Create a test branch and make a change to one of your dbt models. Open a pull request. Within 60 seconds, your agent will profile the model, update the documentation, and commit the changes directly to your PR branch. You can go to the “Runs” tab in the web app to see the full execution log, including which queries the agent ran, what files it modified, and all reasoning steps.