Getting Started with OrKa

1. Install

git clone https://github.com/marcosomma/orka-resoning.git
cd orka
python -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt

2. Configure .env

OPENAI_API_KEY=...
OPENAI_MODEL=gpt-3.5-turbo

3. Run Demo

python test_run.py

This runs orka.yaml against a sample input.

4. Inspect Logs

redis-cli xrevrange orka:memory + - COUNT 5

5. Edit orka.yaml

• Change prompts
• Add agents
• Insert fallback paths
• Configure timeouts and concurrency limits

6. Create Custom Agent

# Modern async pattern (recommended)
from orka.agents.base_agent import BaseAgent

class MyCustomAgent(BaseAgent):
    async def _run_impl(self, ctx):
        input_data = ctx.get("input")
        # Your custom processing logic
        return "processed result"

7. Add to Your YAML Config

- id: my_agent
  type: my_custom_agent
  prompt: Process this input
  queue: orka:my_agent
  timeout: 30.0  # Optional: Set custom timeout
  max_concurrency: 5  # Optional: Set concurrency limit

8. Need Help?

GitHub issues or Discord: https://discord.gg/UthTN8Xu