Tool use (LLM)
Tool use is the umbrella term for any LLM mechanism that lets the model invoke external functions, APIs, or services — function calling, code interpreter, MCP servers, browser actions.
Tool use turned LLMs into actors instead of just generators. The model emits a structured tool call (function name + arguments), the harness executes it, and the result flows back into the conversation. By 2026 every frontier model supports tool use natively; the engineering question is no longer whether but how to design tools well. Best practices: tight schemas, descriptive names, errors readable by the model, tight result payloads, fewer than ~30 tools per agent.
When to use tool use (llm)
- Any agent.
- Workflows where the model needs current data or destructive actions.
Common mistakes
- Too many tools — mis-routing rates climb past ~30.
- Vague tool descriptions — the model uses them as documentation.
- Returning huge JSON dumps — context blows up.
FAQ
What is tool use (llm)?
Tool use is the umbrella term for any LLM mechanism that lets the model invoke external functions, APIs, or services — function calling, code interpreter, MCP servers, browser actions.
When should I use tool use (llm)?
Any agent. Workflows where the model needs current data or destructive actions.
What are the most common mistakes with tool use (llm)?
Too many tools — mis-routing rates climb past ~30. Vague tool descriptions — the model uses them as documentation. Returning huge JSON dumps — context blows up.
Related terms
- Function calling (tool use) — Function calling lets a language model emit a structured request to invoke a developer-defined tool, enabling reliable JSON output and agent workflows.
- AI agent — An AI agent is a system where a language model autonomously plans and executes a sequence of tool calls to accomplish a goal.
- MCP (Model Context Protocol) — MCP is an open protocol from Anthropic that standardises how language models connect to external tools, data sources, and prompts — the USB-C of LLM integrations.
- Code interpreter (LLM tool) — A code interpreter is a sandboxed execution environment that lets a language model run code (usually Python) it generates, inspect the results, and iterate — turning the model into a data analyst.
Last updated: 2026-06-01. Raw markdown: https://promtable.com/glossary/tool-use.md.