Function Calling

The ability of language models to generate structured output that invokes external functions or APIs, enabling interaction with external systems and data.

In Depth

Function calling (also called tool use) is a capability of modern language models that allows them to generate structured output requesting the execution of specific external functions, APIs, or tools rather than just producing text. The model analyzes the user request, determines which function would be helpful, generates the appropriate parameters, and the calling application executes the function and returns the results to the model for further processing.

The function calling workflow operates through a defined protocol. The developer provides the model with a set of available function definitions, including names, descriptions, and parameter schemas. When the model determines that calling a function would help fulfill the user request, it generates a structured function call with the appropriate arguments instead of a text response. The application executes the function, returns the result to the model, and the model uses this information to formulate its final response to the user.

Function calling is supported by major model providers including OpenAI (function calling / tool use), Anthropic (tool use), Google (function calling), and open-source models through frameworks like LangChain. The reliability of function calling depends on the model ability to correctly identify when a function should be called, select the right function from available options, and generate valid parameters. Structured output guarantees and JSON schema validation improve reliability for production applications.

Function calling is the foundation for AI agent architectures, enabling models to interact with databases, APIs, file systems, web services, and enterprise applications. Common use cases include database queries from natural language, API orchestration across multiple services, calculations and data transformations, information retrieval from internal systems, and workflow automation. Production implementations require robust error handling, parameter validation, rate limiting, and access control to ensure safe and reliable function execution.

Related Terms

AI Agent

An autonomous AI system that can perceive its environment, make decisions, use tools, and take actions to accomplish goals with minimal human intervention.

Agentic Workflow

An AI-driven process where language models autonomously plan, execute, and iterate through multi-step tasks using tools, memory, and decision-making.

Large Language Model (LLM)

A neural network with billions of parameters trained on massive text corpora that can understand, generate, and reason about natural language.

Prompt Engineering

The systematic practice of designing and optimizing input prompts to elicit accurate, relevant, and useful outputs from large language models.

Tokens

The fundamental units of text that language models process, representing words, subwords, or characters depending on the tokenization method.

Related Services

Cloud AI Modernisation

Refactoring AWS, Azure, GCP, and Oracle workloads into production-grade AI stacks. Multi-cloud RAG pipelines, observability, guardrails, and MLOps that slot into existing engineering rhythms.

NVIDIA Blueprint Launch Kits

In-a-box deployments for Enterprise Research copilots, Enterprise RAG pipelines, and Video Search & Summarisation agents with interactive Q&A. Blueprints tuned for your data, infra, and compliance profile.

Need Help With Function Calling?

Our team has deep expertise across the AI stack. Let's discuss your project.

Get in Touch