Get from zero to your first multi-agent run in under five minutes.

Prerequisites

  • Python 3.12+
  • Node.js 18+ (web UI only)
  • At least one of: OpenAI API key, Anthropic API key, or Ollama running locally

Step 1 — Clone

git clone https://github.com/zlatko-lakisic/agentic-orchestration.git
cd agentic-orchestration/agentic-orchestration-tool

Step 2 — Setup (automated)

Windows:

.\setup.ps1

macOS / Linux:

chmod +x setup.sh && ./setup.sh

Or manually:

python -m venv .venv
source .venv/bin/activate   # Windows: .venv\Scripts\activate
pip install -r requirements.txt
cp .env.example .env

Step 3 — Configure

Edit .env. Minimum required: set at least one of OPENAI_API_KEY, ANTHROPIC_API_KEY, or configure Ollama (OLLAMA_HOST). Everything else is optional — see the Configuration reference.

Tip: For local-only runs, start Ollama and set OLLAMA_HOST=http://127.0.0.1:11434. The planner can use AGENTIC_PLANNER_MODEL=ollama/llama3.2 once the model is pulled.

Step 4 — First run

Dynamic mode (LLM plans the steps):

python main.py --dynamic "Summarize the top use cases for multi-agent AI systems"

Static workflow (predefined YAML):

python main.py --batch

Router mode (Ollama picks the best workflow):

python main.py "Generate taglines for a developer CLI tool"

Step 5 — Verify your catalog (optional)

Run the platform agent harness to confirm agent YAML is valid and (with API keys) backends are reachable:

# L0 — no API keys; validates full catalog YAML
python main.py --harness-batch --harness-tier static

# L1 — credentialed cloud agents only
python main.py --harness-batch --harness-tier connectivity --harness-filter "gpt_*"

# L2 — single-agent smoke (requires OPENAI_API_KEY or similar)
python main.py --harness-agent gpt_research --harness-tier smoke

See Agent harness roadmap for tiers, profiles, and CI integration.

Step 6 — (Optional) Web UI

cd ../agentic-orchestration-web
npm install
npm start
# Open http://127.0.0.1:3847

Next steps

Topic Link
Browse agent templates Agent Catalog
Verify catalog agents Agent harness
Browse MCP integrations MCP Catalog
Agent skills playbooks Agent Skills
Environment variables Configuration
How components fit together Architecture