Skip to main content
Automations are DAG-based pipelines for scheduled or triggered data processing. They can execute SQL, transform data with AI, and send notifications.
The automations command is aliased as workflows - both work identically.

​
What Automations Can Do

  • Execute SQL queries against Flipside’s data warehouse
  • Transform data with AI/LLM steps
  • Send email notifications
  • Reference uploaded CSV files
  • Accept user inputs via JSON Schema
  • Run on a schedule or be triggered manually

​
Creating Automations

Initialize a new automation: 
flipside automations init my_automation
This creates a YAML file defining the DAG: 
name: my_automation
description: Daily DeFi volume report
inputs:
  type: object
  properties:
    chain:
      type: string
      default: ethereum
steps:
  - id: query_volumes
    type: sql
    sql: |
      SELECT platform, SUM(amount_usd) as volume
      FROM {{inputs.chain}}.defi.ez_dex_swaps
      WHERE block_timestamp >= CURRENT_DATE - 1
      GROUP BY 1
      ORDER BY 2 DESC
  - id: notify
    type: email
    depends_on: [query_volumes]
    to: [email protected]
    subject: Daily DeFi Report

​
Deploying Automations

Push your automation to deploy: 
flipside automations push my_automation.yaml

​
Running Automations

# Run with default inputs
flipside automations run <automation-id>

# Run with custom inputs
flipside automations run <automation-id> -i '{"chain": "arbitrum"}'

​
Automation Commands

CommandDescription
flipside automations listList all automations
flipside automations get <id>Get automation details
flipside automations init <name>Create new automation template
flipside automations push <file.yaml>Deploy automation
flipside automations pull <id>Download automation as YAML
flipside automations run <id>Trigger a run
flipside automations runs <id>List runs for an automation
flipside automations delete <id>Delete an automation

​
Step Types

TypeDescription
sqlExecute a SQL query
llmProcess data with an AI model
emailSend an email notification

​
Using Inputs

Define inputs with JSON Schema: 
inputs:
  type: object
  properties:
    chain:
      type: string
      enum: [ethereum, arbitrum, base]
      default: ethereum
    days:
      type: integer
      default: 7
Reference inputs in SQL with {{inputs.property}}: 
steps:
  - id: query
    type: sql
    sql: |
      SELECT * FROM {{inputs.chain}}.core.fact_blocks
      WHERE block_timestamp >= CURRENT_DATE - {{inputs.days}}

​
Step Dependencies

Use depends_on to control execution order: 
steps:
  - id: step_a
    type: sql
    sql: ...
  - id: step_b
    type: sql
    depends_on: [step_a]
    sql: ...
  - id: step_c
    type: email
    depends_on: [step_a, step_b]
    ...

​
Monitoring Runs

# List runs for an automation
flipside automations runs <automation-id>

# Chat about a specific run
flipside chat --run <run-id>