Agent settings

The Agent Settings page provides essential configuration options for managing your AI agent. This page allows you to define the agent’s purpose, control access settings, enable audit logging, and apply guardrails to ensure appropriate responses.

Agent ID

Each agent has a unique, immutable ID provided when creating the agent. You can copy this ID for use in API calls and in Squid’s SDKs integrations, but it cannot be changed.

Agent Description

You can give your agent a short description to clarify its purpose. Updating this description can help keep track of the agent’s role, especially when managing multiple agents.

Set Agent to Public

By default, agents require authentication, but you have the option to make an agent public. Public agents can be accessed without an API key, making them easier to use in open environments. However, if your agent handles sensitive or proprietary information, be sure to review the documentation on securing public agents before enabling this setting.

Report Audit Logs

Enabling audit logs allows you to track all agent interactions and responses in the Audit Log tab of the Studio. This can be useful for troubleshooting, monitoring agent performance, and maintaining compliance with data policies.

Agent Guardrails

Guardrails help enforce response quality and compliance. You can enable specific restrictions or can define a custom guardrail policy for more tailored restrictions. For more information, visit the guardrails documentation.

Danger Zone (Delete Agent)

If you ever need to remove an agent permanently, you can do so from this section. Deleting an agent is irreversible, so ensure you have backed up any necessary data before proceeding.