<aside> 👉 You decide the level of cognitive firepower your agents need.

</aside>

🧠 SwarmOS Execution Modes: Classic vs. Fusion (Trinidad Mode)

SwarmOS offers two distinct execution layers—each designed for different levels of agent autonomy, complexity, and AI orchestration:

⚙️ Classic Mode (Default)

Run 1 agent → Use 1 brain → Return 1 result.

This is your go-to when:

• You want direct control of a single agent.
• You’re running lightweight tools or skills.
• You don’t need arbitration or fusion.

Usage:

bash
Copy code
python -m swarmos.cli.main run

SwarmOS loads the agents defined in agents.yaml, executes each one using the model of your choice, and prints the result. Fast, simple, surgical.

🧬 Trinidad Mode (Fusion Layer)

Run 1 agent → Use 3 brains → Fuse their minds → Return the best response.

This is parallel intelligence—SwarmOS spins up GPT, Claude, and Grok simultaneously, fuses their responses using your selected strategy (bias, vote, or merge), and gives the agent a fused output.

Usage:

bash
Copy code
python -m swarmos.cli.main run --trinidad-mode

Optional: Set arbitration strategy