Agents and tool use · MLD-1246
Tool use
A model-driven system requesting external functions or services and consuming their results.
Original reference page · detailed expansion pending. Original definitions, examples, notes and citations are preserved below. New examples are illustrative, not benchmark results.
Original example
Call a calculator or document search tool.
Original caveat
The surrounding application executes tools; text alone does not.
Context and aliases
Level: Intermediate. Stages: Inference · Deployment · Evaluation.
Tags: SLM · LLM · RL · Agents.
Aliases: Tool calling.
Related concepts
Concepts linking here
Technology connections
- Agent2Agent (A2A) — Protocol for interoperable communication between agent systems.
- Amazon Bedrock — AWS service for accessing supported foundation models and AI capabilities.
- Anthropic Claude API — Hosted access to Claude models and supported tools.
- Cohere API — Hosted language, embedding and reranking interfaces.
- CrewAI — Framework for role-oriented multi-agent workflows.
- DSPy — Programming and optimization framework for language-model pipelines.
- Fireworks AI — Managed inference and model-customization platform.
- Gemini API — Google's developer interface for supported Gemini models.
- Gemini Enterprise Agent Platform — Google Cloud managed AI platform; current documentation destination for former Vertex AI material.
- Guidance — Library for controlling model generation with structured programs.
- Haystack — Composable pipelines for search and language-model applications.
- Instructor — Library for typed structured responses from supported model APIs.
- JSON Schema — Standard vocabulary for describing and validating JSON documents.
- LangChain — Integrations and abstractions for building model-powered applications.
- LangGraph — Stateful graph-based agent orchestration.
- LiteLLM — Provider-normalizing Python library and proxy gateway.
- LlamaIndex — Data connectors, indexing and retrieval components for LLM applications.
- Microsoft Agent Framework — Framework for building and coordinating agents with enterprise integrations.
- Microsoft Foundry — Microsoft platform for building and operating AI applications and model access.
- Mistral API — Hosted access to Mistral models and supported document/model services.
- Model Context Protocol (MCP) — Protocol for connecting AI applications to tools and context servers.
- OpenAI Agents SDK — SDK for model-driven tools, handoffs and tracing.
- OpenAI Responses API — Hosted interface for model responses and supported built-in tools.
- OpenAPI — Language-neutral description format for HTTP APIs.
- OpenRouter — Unified gateway to multiple model providers.
- Outlines — Library for constraining supported model generation to formats.
- Pydantic — Python type-based data validation and serialization library.
- PydanticAI — Typed agent framework built around Python and Pydantic.
- Semantic Kernel — SDK for integrating models, functions and orchestration into applications.
- smolagents — Lightweight library for agents using model-selected tools and code.
- Temporal — Durable workflow execution platform.
- Together AI — Hosted inference and customization platform for supported models.
- XGrammar — Grammar-constrained decoding engine.
Original references
- Anthropic: Building effective agents ↗ · Reference link
- Model Context Protocol Python SDK: first steps ↗ · Consulted
- ReAct: Synergizing Reasoning and Acting in Language Models ↗ · Reference link
- Original reference URL ↗