Principle:Langchain ai Langchain Tool Binding
| Knowledge Sources | |
|---|---|
| Domains | NLP, Tool_Use, Agentic_AI |
| Last Updated | 2026-02-11 00:00 GMT |
Overview
A configuration step that attaches tool schemas to a chat model so the model can request tool executions as part of its responses.
Description
Tool binding converts tool definitions into the provider's expected format (e.g., OpenAI's tool schema) and attaches them to the model instance. This creates a new Runnable that automatically includes tool schemas in every API request. The model then has the ability to emit structured tool_calls in its response instead of (or in addition to) text content.
Key configuration options include:
- tool_choice: Control whether the model must use a specific tool, any tool, or can choose freely
- strict: Enforce strict JSON schema matching for tool arguments
- parallel_tool_calls: Allow the model to request multiple tools in a single turn
Usage
Bind tools before invoking the model whenever you want the model to be able to call external functions. This is the foundation for building agents.
Theoretical Basis
Tool binding follows the decorator pattern applied to Runnables:
# Abstract algorithm (not real code)
model_with_tools = model.bind(tools=[
{"type": "function", "function": tool_schema}
for tool in tools
])
# model_with_tools.invoke(input) now includes tool schemas in the request
Related Pages
Implemented By
Page Connections
Double-click a node to navigate. Hold to expand connections.
Principle
Implementation
Heuristic
Environment