Tool Use Agent
Augment an LLM with callable external tools — APIs, code interpreters, databases — so it can take actions and retrieve real-time information beyond its training data.
errorWhat Problem Does the Tool Use Agent Pattern Solve?
A standalone language model is limited to knowledge frozen at training time and cannot perform side effects such as searching the web, executing code, or writing to a database. This makes it unreliable for tasks requiring up-to-date information, precise computation, or integration with live systems.
check_circleHow the Tool Use Agent Pattern Works
Define a catalogue of typed tool schemas (name, description, parameter types) and provide them to the model at inference time. When the model determines a tool is needed, it emits a structured tool-call; the host runtime executes the tool and returns results to the model, which incorporates them into its next reasoning step.
Tool Use Agent Pattern Architecture
Rendering diagram...
Implementation by Language
Tool Use Agent Pattern in the Real World
“A lawyer (the LLM) in a courtroom knows the law but needs a paralegal team (the tools) to pull case files, run searches, and retrieve exhibits. The lawyer directs which file to fetch, the paralegal returns it, and the lawyer integrates that information into their argument — the lawyer's intelligence is amplified by the support staff's ability to reach into the real world.”
Frequently Asked Questions
helpWhat is the difference between Tool Use and function calling?
They refer to the same capability. 'Function calling' is the older term used by OpenAI, while 'Tool Use' is the broader term adopted by Anthropic and others. Both describe an LLM generating structured arguments to invoke external functions.
helpHow do I handle tool call errors gracefully in an agent?
Return the error message as the tool result rather than throwing. This lets the LLM see what went wrong and decide whether to retry with different arguments, try a different tool, or inform the user. Never silently swallow errors.