4arXiv cs.AI (Artificial Intelligence)·
Practitioner guide to LangGraph for stateful multi-step business process agents
An arXiv preprint presents a practitioner-oriented guide to LangGraph, a low-level orchestration framework for stateful agentic workflows, with three executable recipes: SQL analytics with repair loops, agentic RAG with evidence gating, and human-in-the-loop policy review with interrupt/checkpoint recovery. The paper positions LangGraph by workflow-complexity fit rather than as a universal default, contrasting it with ReAct loops, schema-first tools, and DSPy for different use cases. The work targets engineers building long-running, auditable business process agents and explains how typed state, conditional routing, retries, and checkpoints become explicit product behavior.
Related guides (2)
Related events (8)
MyAG: Graph-based framework for composable LLM agent system design and analysis
Researchers introduce MyAG, an open-source framework that structures LLM agent systems using three graph abstractions: a component graph, a workflow graph, and a search graph for runtime execution. The separation enables flexible component reuse across different execution strategies, with support for hierarchical composition via recursive system nodes and built-in monitoring and visualization tools. Experiments on representative agent applications demonstrate performance-efficiency tradeoff analysis capabilities.
A Methodology for Selecting and Composing Runtime Architecture Patterns for Production LLM Agents
This paper introduces the stochastic-deterministic boundary (SDB) as a foundational architectural primitive for production LLM agent runtimes, defining it as a four-part contract (proposer, verifier, commit step, reject signal) governing how LLM outputs become system actions. The authors organize agent runtime design around Coordination, State, and Control concerns, presenting a catalog of six runtime patterns applicable to conversational, autonomous, and long-horizon agents. A five-step pattern-selection methodology and diagnostic procedure mapping production failures to pattern weaknesses are contributed, along with a newly named failure mode—replay divergence—where LLM consumers of deterministic event logs produce inconsistent outputs across model versions or prompt changes. The paper argues that as model variance decreases, architectural pattern choice and SDB strength become the dominant reliability levers.
Langflow: Visual AI Agent and Workflow Builder Trending on GitHub
Langflow is an open-source Python framework for building and deploying AI-powered agents and workflows, currently accumulating 148,425 total GitHub stars with 155 new stars today. It provides a visual interface for composing LLM-based pipelines and agent workflows. The continued traction signals ongoing community interest in low-code/visual tooling for AI agent construction.
Open-source LLMs as LangChain Agents
This Hugging Face blog post explores using open-source LLMs as agents within the LangChain framework. It examines the capability of various open-weight models to perform tool use, reasoning, and multi-step task execution in agentic settings. The post likely benchmarks or compares several models on agent-relevant tasks, providing practical guidance for deploying open-source alternatives to proprietary models in agent pipelines.
GaP: Graph-as-Policy multi-agent harness for reliable robot automation via self-learning computation graphs
Researchers introduce Graph-as-Policy (GaP), a multi-agent coding harness that generates directed computation graphs combining perception, planning, and control nodes for robotic 'Variational Automation' tasks — those with high variation in object geometry and pose. GaP uses an internal simulation environment to rehearse and iteratively refine graph structures in parallel, improving success rates without relying solely on model-free policies. Evaluation across 8 new benchmarks (4 simulated, 4 real-world) shows significant outperformance over baselines. The work bridges agentic coding systems with interpretable robot programming and Task and Motion Planning.
HULAT2 multi-agent LangGraph system for Spanish Easy-to-Read text simplification at MER-TRANS 2026
Researchers from HULAT2-UC3M describe their submission to the MER-TRANS 2026 shared task on multilingual Easy-to-Read translation, using a LangGraph-based multi-agent workflow combining Gemini 2.5 Flash and RigoChat-7B-v2. The best run (RUN1) achieved a SARI score of 44.05 using Event-Condition-Action routing and internal quality signals, outperforming a LoRA-adapted generate-evaluate-regenerate baseline. Results show signal-guided multi-agent routing outperforms linear regeneration, while adding lexical support did not automatically improve reference-based scores.
Tutorial paper surveys agentic LLM systems from research to production deployment
A tutorial paper from arXiv surveys the transition of LLM-based agentic systems from research prototypes to production deployments across software engineering, scientific discovery, and finance. The work covers reasoning and planning, multi-agent coordination, and evaluation, with applied case studies in pharmaceutical discovery and financial systems. It identifies common design patterns for successful deployments and practical mitigation strategies for failure modes including verification pipelines, fallback mechanisms, and human-in-the-loop supervision.
GRACE: Graph-Regularized Agentic Context Evolution for reliable long-horizon instruction updates
Researchers introduce GRACE, a method that maintains a deployed LLM agent's persistent system-level instructions as a typed semantic graph rather than flat text, enabling local verification of updates within typed node neighborhoods. Evaluated on a telecom agent harness derived from τ²-bench under distribution shift, GRACE improves pass³ reliability from 0.091 (Gemini 2.5 Flash zero-shot) to 0.673±0.136, surpassing a Gemini 3.1 Pro zero-shot reference of 0.242. The work identifies structural substrate and consolidation mechanisms as key requirements for reliable long-horizon agentic context evolution. The flat-text baseline finishes at 0.191, underscoring the practical gap GRACE addresses.