How langchain-ai/langgraph-101 works
LangGraph 101 is a hands-on learning repository: learners progress from agent fundamentals to advanced workflow patterns, with standalone implementations available to run in LangGraph Studio. Its examples cover tool-using agents, email triage, supervised specialist agents, parallel research, and Deep Agents capabilities. The sampled source confirms several runtime connections; for unsampled implementations, edges rely only on README-documented relationships. Notebooks are represented as learning experiences rather than an invented execution pipeline.
Main entry points: deep_agent/agent.py Deep research agent · memory_enabled_music_store_supervisor_with_interrupt.py Verified memory workflow · models.py Model configuration · 101/agent.py Weather assistant
Components 24
Learning tracks
- Agent fundamentals
notebooks/101/101_langchain_langgraph.ipynb - Middleware and HITL
notebooks/101/102_middleware.ipynb - Email agent lesson
notebooks/201/email_agent.ipynb - Multi-agent lesson
notebooks/201/multi_agent.ipynb - Research lesson
notebooks/201/research_agent.ipynb - Deep Agents lesson
notebooks/201/deep_agents.ipynb
Standalone agents
- Weather assistant
agents/101/agent.py - Email triage agent
agents/email_agent/graph.py - Music store supervisor
agents/music_store/music_store_supervisor.py - Music catalog agent
agents/music_store/music_agent.py - Invoice agent
agents/music_store/invoice_agent.py - Verified memory workflow
agents/music_store/memory_enabled_music_store_supervisor_with_interrupt.py - Parallel research agent
agents/researcher/graph.py - Research prompts
agents/researcher/prompts.py - Deep research agent
agents/deep_agent/agent.py - On-demand writing skills
agents/deep_agent/skills - Agent instructions
agents/deep_agent/AGENTS.md
Shared capabilities
- Model configuration
utils/models.py - Chinook database setup
utils/utils.py
Tools and integrations
- Email and calendar tools
mcp/email_tools.py - Open-Meteo
- Tavily search
- Long-term memory store
Other components
- Learner
Connections 28
- Learner → Agent fundamentals: studies
- Learner → Middleware and HITL: studies
- Learner → Email agent lesson: studies
- Learner → Multi-agent lesson: studies
- Learner → Research lesson: studies
- Learner → Deep Agents lesson: studies
- Learner → Weather assistant: runs
- Learner → Email triage agent: runs
- Learner → Verified memory workflow: runs
- Learner → Parallel research agent: runs
- Learner → Deep research agent: runs
- Weather assistant → Model configuration: invokes
- Weather assistant → Open-Meteo: requests weather
- Email triage agent → Model configuration: invokes
- Verified memory workflow → Music store supervisor: delegates
- Music store supervisor → Music catalog agent: routes music queries
- Music store supervisor → Invoice agent: routes invoice queries
- Verified memory workflow → Chinook database setup: loads customer data
- Music catalog agent → Chinook database setup: queries catalog
- Invoice agent → Chinook database setup: queries invoices
- Verified memory workflow → Long-term memory store: reads and writes memory
- Parallel research agent → Research prompts: uses prompts
- Parallel research agent → Model configuration: invokes
- Deep research agent → Model configuration: invokes
- Deep research agent → Tavily search: searches web
- Deep research agent → On-demand writing skills: loads skills
- Deep research agent → Agent instructions: loads instructions
- Deep research agent → Long-term memory store: stores memories