Tools
Tools are extensions that enable Large Language Models (LLMs) to interact with external systems, retrieve data, or perform specific actions beyond the original generation. They expand an agent's capabilities by allowing it to execute functions, query data sources, call APIs, or access other resources in response to user requests.
When an LLM determines that a tool is needed to fulfill a request, it can invoke one or more of the configured tools with the necessary parameters. The tool executes its function and returns results to the LLM, which then incorporates this information into its response. This allows agents to provide dynamic, real-time information and perform actions rather than relying solely on the training data.
The agent model provides two levels at which tools can be configured:
- Operation-level tools: tools can be added to specific operations within an agent model. This approach gives you granular control, enabling particular capabilities only where they're needed.
- Agent-level tools: tools can also be configured at the model level, making them available to all operations within that agent. This is useful for tools that should be universally accessible.
Not all LLM models support tool-calling, make sure to check the model documentation of your LLM provider.