Pre-built AINL workflow templates for common use cases.

Creates a REST API endpoint that fetches and processes data.

Use when:

• Building API services
• Creating data fetch endpoints
• Need authenticated API calls

Frame variables:

• api_key: Authentication token
• base_url: API base URL

Checks a service health endpoint and alerts if unhealthy.

Use when:

• Monitoring service health
• Need recurring checks
• Want automated alerts

Frame variables:

• health_url: Health check endpoint
• alert_webhook: Where to send alerts

Schedule: Every 5 minutes (*/5 * * * *)

ETL workflow: Extract from API, Transform, Load to warehouse.

Use when:

• Building data pipelines
• Daily data exports
• Need scheduled ETL

Frame variables:

• source_api: Data source URL
• warehouse_url: Data warehouse URL
• api_key: API authentication

Schedule: Daily at 2 AM (0 2 * * *)

Monitors Solana wallet balance and alerts if below threshold.

Use when:

• Monitoring crypto wallets
• Need balance alerts
• Working with Solana

Frame variables:

• wallet_address: Solana wallet address
• threshold_sol: Alert threshold in SOL
• alert_webhook: Where to send alerts

Schedule: Hourly (0 * * * *)

AI-powered content moderation workflow.

Use when:

• Content moderation
• AI classification tasks
• Need LLM in workflows

Frame variables:

• content: Content to moderate
• webhook_url: Where to send flags

Requirements: AINL_CONFIG with LLM provider setup

Approval workflow with conditional logic.

Use when:

• Multi-step processes
• Approval workflows
• Conditional automation

Frame variables:

• request_id: Request to process
• api_base_url: API base URL
• approval_webhook: Manual approval URL

# Copy template
cp templates/ainl/monitor_workflow.ainl my_monitor.ainl

# Edit frame variables and logic
nano my_monitor.ainl

# Validate
ainl validate my_monitor.ainl --strict

# Run
ainl run my_monitor.ainl --frame '{"health_url":"https://api.example.com/health","alert_webhook":"https://hooks.slack.com/..."}'

"Create a health monitor based on the monitor_workflow.ainl template for https://api.example.com/health"

Claude will customize the template for your use case.

All templates follow this structure:

# Title and description
# Frame variable declarations (# frame: name: type)

workflow_name [@cron "schedule" | @api "/path"]:
  in: parameters
  
  # Steps
  result = adapter.operation args
  
  # Conditional logic (if needed)
  if condition:
    # then branch
  
  # Output
  out {result: value}

Templates use frame variables for configuration. These are passed at runtime:

// In ainl_run MCP call
{
  source: "...",
  frame: {
    api_key: "sk-...",
    webhook_url: "https://...",
    threshold: 100
  }
}

Declare frame variables in comments:

# frame: api_key: string
# frame: threshold: number

Templates use these adapters (enable in ainl_run):

• http: api_endpoint.ainl, monitor_workflow.ainl, data_pipeline.ainl, multi_step_automation.ainl
• solana: blockchain_monitor.ainl
• llm: llm_workflow.ainl (requires config)
• core: All templates (always available)

Example ainl_run call:

ainl_run({
  source: "...",
  frame: {...},
  adapters: {
    enable: ["http"],
    http: {
      allow_hosts: ["api.example.com"],
      timeout_s: 30
    }
  }
})

1. Always validate first: ainl validate template.ainl --strict
2. Test with sample data: Use ainl_run with test frame variables
3. Start simple: Copy a template, test it, then customize
4. Use frame hints: Add # frame: comments for all variables
5. Set timeouts: Specify timeout for HTTP calls (3rd argument)
6. Handle errors: Use if conditions to check for failures

Using these templates for recurring tasks:

Compile once, run many times = massive token savings!

To add a new template:

1. Create .ainl file following the structure
2. Add frame variable comments
3. Test with ainl validate and ainl run
4. Document in this README
5. Submit PR

• AINL Language Guide: ../docs/AINL_LANGUAGE_GUIDE.md
• MCP Tools: Use ainl_capabilities to see available adapters
• AINL Spec: https://ainativelang.com/docs