Skip to main content

getting-started

Getting Started with IOA Core

Welcome to IOA Core! This guide will help you get up and running with the IOA (Intelligent Orchestration Architecture) framework in minutes.

What is IOA Core?​

IOA Core is the open-source framework for governed AI orchestration — bringing verifiable policy, evidence, and trust to every workflow.

Key capabilities:

  • Immutable Audit Chains: Cryptographically verified audit logging
  • System Laws Framework: Seven governing principles for AI orchestration
  • 6 LLM Providers: OpenAI, Anthropic, Gemini, DeepSeek, XAI, and Ollama
  • Vendor-Neutral Quorum: Multi-agent consensus with graceful scaling
  • Memory Fabric: Multi-tier storage with hot and cold layers
  • Working Examples: Six complete examples to get you started

System Requirements​

  • Python 3.10 or higher
  • pip or poetry
  • (Optional) LLM provider API keys for live testing

Installation​

Quick Install​

Install IOA Core from PyPI:

pip install ioa-core

Development Install​

For development, install from source:

git clone https://github.com/orchintel/ioa-core.git
cd ioa-core
pip install -e ".[dev]"

Verify Installation​

Check that IOA Core is properly installed:

# Check Python environment
python --version  # Should be 3.10+

# Run system health check
python examples/30_doctor/doctor_check.py

Expected output:

{
  "python_version_ok": true,
  "python_version": "3.10.0",
  "overall_health": "healthy"
}

Quick Start​

1. Bootstrap a Project​

Create a new IOA project:

python examples/00_bootstrap/boot_project.py my-ai-system

This creates a project directory with:

  • my-ai-system/ioa.yaml - Configuration file
  • my-ai-system/README.md - Project documentation

2. Run Your First Workflow​

Execute a governed workflow:

python examples/10_workflows/run_workflow.py

Output shows:

  • Task execution
  • Governance policy applied
  • Audit evidence ID
  • System Laws enforced

3. Test Provider Connectivity​

Check provider status:

# Mock provider (no API key needed)
IOA_PROVIDER=mock python examples/40_providers/provider_smoketest.py

# Real provider (requires API key)
IOA_LIVE=1 IOA_PROVIDER=openai python examples/40_providers/provider_smoketest.py

4. Run Multi-Agent Quorum​

Demonstrate vendor-neutral roundtable:

python examples/20_roundtable/roundtable_quorum.py "Analyze this code (ok)"

Features:

  • 3-agent quorum voting
  • 2-of-3 approval threshold
  • Vendor-neutral consensus

Examples​

IOA Core includes six working examples:

1. Bootstrap​

File: examples/00_bootstrap/boot_project.py

Create a new IOA project with configuration and schemas.

python examples/00_bootstrap/boot_project.py my-ai-system

2. Workflows​

File: examples/10_workflows/run_workflow.py

Run governed workflows with policy enforcement and audit logging.

python examples/10_workflows/run_workflow.py

3. Roundtable​

File: examples/20_roundtable/roundtable_quorum.py

Multi-agent consensus with vendor-neutral quorum policy.

python examples/20_roundtable/roundtable_quorum.py "Your task (ok)"

4. Doctor​

File: examples/30_doctor/doctor_check.py

System health check and environment validation.

python examples/30_doctor/doctor_check.py

5. Providers​

File: examples/40_providers/provider_smoketest.py

Test LLM provider connectivity and configuration.

IOA_PROVIDER=mock python examples/40_providers/provider_smoketest.py

6. Ollama Turbo​

File: examples/50_ollama/turbo_mode_demo.py

Local model optimization with turbo mode.

python examples/50_ollama/turbo_mode_demo.py turbo_cloud

API Keys Setup (Optional)​

For live testing with real LLM providers, set up your API keys:

# Set API keys
export OPENAI_API_KEY="sk-your-openai-key"
export ANTHROPIC_API_KEY="sk-ant-your-anthropic-key"
export XAI_API_KEY="xai-your-xai-key"
export GOOGLE_API_KEY="your-google-key"
export DEEPSEEK_API_KEY="your-deepseek-key"

# Enable live mode
IOA_LIVE=1 python examples/10_workflows/run_workflow.py

What You Get​

IOA Core v2.5.2 provides:

  • ✅ Working governance framework
  • ✅ Immutable audit chains with cryptographic verification
  • ✅ System Laws enforcement at runtime
  • ✅ 6 LLM provider support with unified interface
  • ✅ Memory fabric (hot & cold storage)
  • ✅ Vendor-neutral quorum policy for multi-agent consensus
  • ✅ Comprehensive examples and documentation
  • ✅ CLI interface for system management
  • ✅ Production-ready architecture

Next Steps​

Getting Help​

Resources​

Ready to start building? Check out our Quick Start Guide to get up and running in minutes.