Getting Started

The shortest path into Operon if you care about practical value first.

Installation

pip install operon-ai

For provider-backed workflows, configure the model backend you want through the existing Nucleus provider layer.

Recommended First Steps

1. Start with the Pattern-First API.
2. Run the pattern-first example and the skill-organism example.
3. Use the Examples page to find the shortest path into the rest of the library.

First Practical Questions

If you are evaluating Operon, the right first questions are usually:

• should this stay single-agent?
• does it need a reviewer bottleneck?
• is it actually decomposable into specialists?
• should fixed work go to a cheap model and fuzzy work to a stronger one?

Minimal Entry Points

• advise_topology(...) for architecture guidance
• reviewer_gate(...) for one-worker-plus-reviewer
• specialist_swarm(...) for centralized specialist decomposition
• skill_organism(...) for multi-stage workflows with attachable components