Skip to main content

Menu

Home
GitHub(새 탭에서 열림)

Overview

Quickstart
Customization
Models
Comparison with Claude Agent SDK
Changelog

Going to production
Fault tolerance

Tools
Backends
Permissions
Multimodality
Sandboxes
Interpreters
Event streaming
Streaming

Skills
Memory
Retrieval
Context engineering
Profiles
OpenWiki

Subagents
Dynamic subagents
Async subagents

Human-in-the-loop
Grading rubrics

Overview
Prebuilt middleware
Custom middleware

Overview

Subagent streaming
Todo list
Sandbox

Agent Client Protocol (ACP)
MCP with LangChain
A2A with LangSmith
AG-UI

Overview

Install
Quickstart
Changelog
Philosophy

Agents
Models
Messages
Tools
Short-term memory
Event streaming
Streaming
Structured output

Overview
Prebuilt middleware
Custom middleware

Overview

Markdown messages
Tool calling
Headless tools
Human-in-the-Loop
Branching chat
Reasoning tokens
Structured output
Message queues
Join & rejoin streams
Time travel

Overview
Controlled
Declarative
Open-ended

Overview
CopilotKit
AI Elements
assistant-ui
OpenUI

Guardrails
Runtime
Context engineering
Human-in-the-loop
Retrieval
Long-term memory

Overview
Tools
Connections
Authentication

Overview
Subagents
Handoffs
Skills
Router
Custom workflow

LangSmith Studio
Agent Chat UI

Overview
Unit testing
Integration testing
Agent Evals

Deployment
Observability

Overview
Studio

Install
Quickstart
Local server
Changelog
Thinking in LangGraph
Workflows + agents

Persistence
Checkpointers
Stores
Fault tolerance
Event streaming
Streaming
Interrupts
Time travel
Memory
Subgraphs

Application structure
Test
Backward compatibility
LangSmith Studio
Agent Chat UI
Deployment
LangSmith Observability

Overview
Graph execution
Custom stream channels

Runtime

Choosing APIs
Graph API
Use the graph API

Functional API
Use the Functional API

Overview
Quickstart
Integrations
Search and link
Visualize
CLI reference
Customize OpenWiki
Providers
Automate updates
Changelog

Code mode
Personal mode

LangChain integrations
All providers

OpenAI
Anthropic
Google
AWS
NVIDIA
Hugging Face
Microsoft
Ollama
Groq
Fireworks

Chat models
Tools and toolkits
Middleware
Sandboxes
Backends
Checkpointers
Stores
Retrievers
Text splitters
Embedding models
Vector stores
Document loaders

Learn

Data Analysis
Deep Research
Content Builder
RAG

Deep Agent
Semantic search
SQL agent
Voice agent

Subagents: Personal assistant
Handoffs: Customer support
Router: Knowledge base
Skills: SQL assistant

Custom RAG agent
Custom SQL agent

LangChain vs. LangGraph vs. Deep Agents
Providers and models
Component architecture
Memory
Context
Human-in-the-loop
Graph API
Functional API

LangChain Academy
Case studies
Get help

Overview

Deep Agents
LangChain SDK
LangGraph SDK
Integrations
Errors

Versioning
Changelog

LangChain v1
LangGraph v1

LangChain v1
LangGraph v1
langgraph-supervisor
langchain-mcp-adapters

Release policy
Security

Overview

Documentation
Code

Guide
Implement
Standard tests
Publish
Co-marketing

Overview

Quickstart
Customization
Models
Comparison with Claude Agent SDK
Changelog

Going to production
Fault tolerance

Tools
Backends
Permissions
Multimodality
Sandboxes
Interpreters
Event streaming
Streaming

Skills
Memory
Retrieval
Context engineering
Profiles
OpenWiki

Subagents
Dynamic subagents
Async subagents

Human-in-the-loop

Overview
Prebuilt middleware
Custom middleware

Overview

Subagent streaming
Todo list
Sandbox

Agent Client Protocol (ACP)
MCP with LangChain
A2A with LangSmith
AG-UI

Overview

Install
Quickstart
Changelog
Philosophy

Agents
Models
Messages
Tools
Short-term memory
Event streaming
Streaming
Structured output

Overview
Prebuilt middleware
Custom middleware

Overview

Markdown messages
Tool calling
Headless tools
Human-in-the-Loop
Branching chat
Reasoning tokens
Structured output
Message queues
Join & rejoin streams
Time travel

Overview
Controlled
Declarative
Open-ended

Overview
CopilotKit
AI Elements
assistant-ui
OpenUI

Guardrails
Runtime
Context engineering
Human-in-the-loop
Retrieval
Long-term memory

Overview
Tools
Connections
Authentication

Overview
Subagents
Handoffs
Skills
Router
Custom workflow

LangSmith Studio
Agent Chat UI

Overview
Unit testing
Integration testing
Agent Evals

Deployment
Observability

Overview
Studio

Install
Quickstart
Local server
Changelog
Thinking in LangGraph
Workflows + agents

Persistence
Fault tolerance
Checkpointers
Stores
Event streaming
Streaming
Interrupts
Time travel
Memory
Subgraphs

Application structure
Test
Backward compatibility
LangSmith Studio
Agent Chat UI
Deployment
LangSmith Observability

Overview
Graph execution
Custom stream channels

Runtime

Choosing APIs
Graph API
Use the graph API

Functional API
Use the Functional API

Overview
Quickstart
Integrations
Search and link
Visualize
CLI reference
Customize OpenWiki
Providers
Automate updates
Changelog

Code mode
Personal mode

Overview
All providers

Overview
Chat
Embeddings
Tools

Overview
Chat
Tools

Overview
Chat
Tools
ChatGoogleGenerativeAI (Legacy)
ChatVertexAI (Legacy)

Overview
ChatBedrockConverse integration
Embeddings

Overview
Chat
Embeddings

Chat models
Tools and Toolkits
LLMs
Middleware
Sandboxes
Backends
Key-value stores
Stores
Document transformers
Model caches

Retrievers
Text splitters
Embedding models
Vector stores
Document loaders
Key-value stores

Learn

Deep Research
Content Builder
RAG

Semantic search
SQL agent
Voice agent

Subagents: Personal assistant
Handoffs: Customer support
Router: Knowledge base
Skills: SQL assistant

Custom RAG agent

LangChain vs. LangGraph vs. Deep Agents
Providers and models
Component architecture
Memory
Context
Human-in-the-loop
Graph API
Functional API

LangChain Academy

Case studies
Get help

Overview

Deep Agents
LangChain SDK
LangGraph SDK
Integrations
Errors

Versioning
Changelog

LangChain v1
LangGraph v1

LangChain v1
LangGraph v1
MCP adapters v2

Release policy
Security

Overview

Documentation
Code

Guide
Implement
Standard tests
Publish
Co-marketing

Overview
Quickstart
CLI reference
Approval modes
Goals and rubrics
Plugins
Python extensions
Memory and skills
Remote sandboxes
Subagents
Model providers
Changelog

Configuration
Credentials
config.toml
Hooks
MCP tools

本页还没有译文

本页 1 段还没译成한국어，下面显示的是英文原文。

Model Context Protocol

还没翻译译自英文（docs.langchain.com）本页 1 段都还没译

AI 翻译·查看英文原文

Adapted from docs.langchain.com/oss/javascript/deepagents/mcp by LangChain, licensed under MIT. Changes were made: this page is re-rendered from the original's source.

이전Agent Client Protocol (ACP)다음A2A with LangSmith

번역 © 2026 fanyi.dev · 원문 © 2026 LangChain (MIT)

본문 넓게 보기