Skip to main content
Tools let your agents take action in the real world. When the LLM decides a tool should be called, Agntix executes it and feeds the result back into the conversation. There are two tool types:

Create a tool

Define a new API or function tool.

List tools

Browse all tools in your organization.

Attach to an agent

Link a tool to an agent using PATCH /v2/chat/agents/.

Test a tool

Use the dashboard playground to test tools before attaching them.

Key concepts

  • name — the function name the LLM sees. Use UPPER_SNAKE_CASE for clarity (e.g. GET_ACCOUNT_BALANCE). Must be alphanumeric with underscores, no leading/trailing underscores.
  • description — tells the LLM when to call this tool. Be specific. The LLM uses this to decide whether to invoke the tool in a given context.
  • ajvPropertiesSchema — a JSON Schema object defining the tool’s input parameters. The LLM constructs its call arguments based on this schema.
  • preExecutionMessage — an optional message spoken/sent to the user while the tool runs (e.g. “Let me look that up for you…”).

API tool example

Tools and agents

A tool must be attached to an agent before it becomes available in conversations. Attach tools via the tools array in Create Agent or Update Agent.