Set up your AI model
Last updated Mar 30, 2026
Every agent needs a language model to reason and act. You select the model when you create an agent, and you can change it from the agent's settings at any time.
When your agent runs, your prompts and task context are sent to your chosen model provider's API. This is subject to that provider's terms of service and privacy policy. Review your provider's documentation before storing sensitive data in agent tasks.
Select a model for an agent
Cloud agents
- Open the agent's edit view in the desktop app.
- Under Provider, choose your LLM provider — for example, OpenAI or Anthropic.
- Under Model, select the specific model you want the agent to use.
- Save the agent.
Available providers and models reflect what your workspace administrator has configured in MultiClaw Cloud.
Local agents
For local agents, available models are determined by your OpenClaw configuration. The model dropdown in the agent builder lists the models your local OpenClaw instance has access to.
Write an effective personality
The model you choose affects the agent's capability, but the personality prompt shapes how it uses that capability.
- A clear, role-focused personality gets more consistent results than a generic one.
- If you need high accuracy for complex tasks, use a more capable model and pair it with a specific personality.