#LangGraph
64 related articles
LangGraph for Beginners: Agent Flow Control and Enterprise Development in Practice
Based on a Bilibili tutorial, this article explains how LangGraph solves the control problem of Agents in fixed flows, compares the hierarchy of LangChain and Deep Agent, and offers scenario selection advice.
Self-Hosting LangGraph: Multi-Worker Coordination and Crash Recovery Challenges
How to design leases, heartbeats, fencing tokens, and crash recovery for self-hosted LangGraph multi-worker setups to prevent stale checkpoint writes.
LangGraph in Practice: Three Parallel Agents per Request at Just $0.22
protoAgent, built on LangGraph, dispatches one request to three parallel agents, each running Claude Code, with a live org chart and metrics—all for $0.22, MIT-licensed and local-first.
LangGraph in Practice: A Deep Dive into Real Developer Workflows and Tech Stacks
Starting from a Reddit developer question, this article analyzes LangGraph's real-world production workflows, tech stack combinations, graph structure patterns, and debugging pain points, summarizing key practices and observability needs for moving from tutorials to production.
Cosine Similarity Isn't Enough: rag-critic, an Adversarial Fact Auditor for LangGraph
Cosine similarity above 0.85 still produces confident hallucinations. rag-critic uses atomic proposition decomposition plus NLI entailment judgment for adversarial fact auditing in LangGraph, blocking and issuing rewrite instructions when the grounding score falls below threshold.
LangGraph Crash Course: Building Enterprise-Grade AI Agents and Workflows from Scratch
LangGraph crash course: learn its independence from LangChain, MIT license, FastAPI high-concurrency deployment, human approval, and memory storage for enterprise AI agents.
LangGraph Retry Safety in Practice: Idempotency, Approval Gates, and Compensation Handlers
Drawing on a Reddit workshop, we break down three LangGraph production patterns: idempotent state machines to avoid duplicate side effects on retry, human approval gates for high-impact decisions, and compensation handlers to undo completed steps—plus small-encoder routing and Wilson interval evaluation.
LangGraph Interrupt Mechanism Explained: Node Re-runs and Approval Payload Safety
Deep dive into LangGraph's interrupt mechanism: interrupt() pauses by raising on first call, and on resume the whole node re-runs, so approval payloads may be regenerated. We explore duplicate side effects, index matching, and solutions like predecessor nodes and the Functional API to keep approval flows safe.
Building AI Agents: A Progressive Practice from Harness to LangGraph
Explore a progressive approach to building AI agents: start with the deepagents harness, add context, skills, and tools, fine-tune with middleware, and use LangGraph workflows when necessary. A practical guide for developers.
LangGraph Agent Architecture Explained: From Agent to Orchestration in Practice
An in-depth analysis of LangGraph's agent architecture: starting from the process control dilemma of LangChain Agents, dissecting LangGraph's low-level orchestration capabilities, core components, and its hierarchical relationship with LangChain and Deep Agents, with practical selection advice.
Building a Medical Graph Engine with FalkorDB+LangGraph to Solve the $260 Billion Claims Dilemma
Doctors spend 2 hours on EHR entry for every 1 hour of care; U.S. hospitals lose $260B yearly to claim denials. How FalkorDB knowledge graphs and LangGraph multi-agents replace vector RAG for deterministic medical necessity proof, NCCI unbundling interception, and cross-encounter contraindication detection.
LangGraph in Practice: From Building Blocks to Production-Grade AI Agents
*LangGraph in Small Pieces* is a hands-on handbook from foundational building blocks to production-grade AI Agents, covering 31 chapters on state, nodes, checkpoints, tool calling, RAG, memory, and evaluation, using local models by default with no API costs.
LangGraph Four-Way Agent Showdown: Plan-and-Execute Costs 2.3× More With No Advantage
Pre-registered LangGraph comparison of four architectures: under the same model, tools, and budget, ReAct, Plan-and-Execute, Supervisor, and Draft-Verify show no significant accuracy difference on 240 legal Q&A tasks, yet Plan-and-Execute costs 2.3× more with no benefit.
Building an AI Directory Content Pipeline with LangGraph and Contentful
Analyzing Monogram's AI directory content pipeline for Quicknode: a 5-node LangGraph workflow combining Tavily retrieval, GPT-4o generation, and Contentful category validation for 60+ structured fields.
Getting Started with LangGraph: From Agent Pain Points to Low-Level Orchestration Principles
A deep dive into LangGraph's low-level orchestration principles, comparing the hierarchy and use cases of LangChain Agent and Deep Agent to explain why fixed-workflow tasks need LangGraph rather than relying solely on Agents.
OpenUI 1.0: Letting LangGraph Agents Generate Interactive Interfaces Directly
OpenUI 1.0 is released: an MIT open-source, model-agnostic generative UI framework with LangChain/LangGraph integration, letting AI agents generate interactive interfaces directly.
Why Is LangGraph Supervisor So Token-Hungry? A Hop-by-Hop Cost Breakdown
A deep dive into why the LangGraph Supervisor pattern burns so many tokens: it's re-invoked after every worker, and context inflates hop by hop. We analyze cost structure via langgraph-supervisor source code and offer five verifiable optimization directions plus hop-by-hop measurement methods.
LangGraph + FastAPI + SSE: A Practical Guide to Building Streaming AI Agents
A hands-on tutorial on building streaming AI agents with LangGraph, FastAPI, and SSE: covering separation of agent orchestration and API transport, Open-Meteo tool integration, SSE streaming responses, and production gaps in persistence, idempotency, and observability.
LangChain or LangGraph? Let create_age…
LangChain or LangGraph? Let create_agent's Return Value Decide
In LangChain 1.4.3, create_agent actually returns a LangGraph graph. Five runnable scripts compare create_agent, StateGraph, and agent-as-node, with selection rules and common error fixes.
Building a Golden Hour Walk Planner with a 1B Model: Let AI Handle Language, Let Code Handle Math
A local 1B-model walk planner that keeps models on language and code on computation, using LangGraph pipelines and double-bracket place name validation to fight hallucinations. Includes qwen2.5 vs llama3.2 benchmarks, all under 0.5GB RAM.