Skip to main content

Quick Start

Get up and running with ShedBoxAI in just a few minutes. This guide will walk you through creating your first AI-powered data processing pipeline with simple YAML configuration.

Overview​

ShedBoxAI transforms complex data processing workflows into simple YAML configurations. Just two steps to get started:

  1. Create a YAML configuration file
  2. Run shedboxai run config.yaml

The framework handles everything else: data connections, transformations, AI integration, and output formatting.

AI-Powered Configuration

Want to generate configurations with Claude Code or other AI assistants? Download the AI Assistant Guide - it contains all the syntax patterns and examples needed for AI to create perfect ShedBoxAI configs for you.

Your First Pipeline​

Let's create a simple pipeline that processes user data and generates insights.

Step 1: Create Sample Data​

Create a file called data/users.csv:

name,age,city,status
John Doe,30,New York,active
Jane Smith,25,Los Angeles,active
Bob Johnson,45,Chicago,inactive
Alice Brown,35,Miami,active

Step 2: Create Configuration​

Create config.yaml:

# Data sources configuration
data_sources:
  users:
    type: csv
    path: data/users.csv

# Processing configuration
processing:
  contextual_filtering:
    users:
      - field: age
        condition: "> 18"
      - field: status
        condition: "== 'active'"
        new_name: "adult_active_users"
  
  format_conversion:
    adult_active_users:
      template: |
        **{{item.name}}** ({{item.age}} years)
        - Location: {{item.city}}
        - Status: {{upper(item.status)}}

# Output configuration  
output:
  type: file
  path: output/results.json
  format: json

Step 3: Run Your Pipeline​

shedboxai run config.yaml

Expected Output​

The pipeline will:

  1. Load user data from CSV
  2. Filter for active users over 18
  3. Format each user with a template
  4. Save results to output/results.json

Key Concepts​

Data Sources​

Connect to various data sources:

  • CSV/JSON/YAML files
  • REST APIs
  • Text files
  • Inline data

Operations​

Transform your data with 6 operation types:

  • Contextual Filtering - Filter data with expressions
  • Format Conversion - Transform with templates
  • Content Summarization - Statistical analysis
  • Relationship Highlighting - Link and derive data
  • Advanced Operations - Group, aggregate, sort
  • Template Matching - Jinja2 templating

Expression Engine​

Use built-in functions in conditions and templates:

condition: "> avg(map(users, 'age'))"
template: "{{upper(concat(item.first_name, ' ', item.last_name))}}"

Next Steps​

Advanced Features​

Need Help?​