AI Agents Blog
Deep dives and practical articles about AI agents — agent frameworks, orchestration patterns, MCP integrations, and autonomous workflows. Curated from our main blog: every card links to the full article. 840 articles and growing.
Codex for Builders: Your AI-Powered Coding Companion
Unlock faster, higher-quality development with Codex, OpenAI's agentic AI teammate. Explore its integrations, powerful GPT-5-Codex model, and practical workflows to supercharge your coding productivity across terminals, IDEs, and more.
MCP for Builders: The Universal Adapter for AI-Powered Apps
Unlock seamless integration of real-world tools into your AI apps using MCP, a single simple interface that acts as a universal adapter. Reduce custom code, speed up development, and scale effortlessly with reusable commands across services.
Claude + LangGraph: Persistent Memory Agents for Complex Workflows
Unlock persistent memory agents with Claude AI and LangGraph for flawless complex workflows. Features checkpointing, error recovery, and multi-step reasoning in Python examples.
Advanced MCP Servers: Extend Claude with Real-Time Database Tools
Supercharge Claude AI with custom MCP servers for real-time Postgres and Redis access. This tutorial walks you through building, Dockerizing, and deploying them with tool-calling prompts.
Claude for Project Management: Auto-Generate Roadmaps, Tasks, and Jira Tickets
Struggling with project roadmaps and Jira tickets? Use Claude AI to auto-generate them instantly, complete with risk analysis and resource allocation—saving hours on PM grunt work.
Claude Messages API Mastery: Building Persistent Memory Chatbots in Python
Build stateful chatbots with Claude's Messages API that remember conversations, profile users, and scale for customer service. Master Python implementation from basics to deployment in this hands-on g
Building a Smart AI Agent with the OpenAI Node.js SDK
Learn to create an intelligent agent that detects your location, checks the weather, and recommends local activities using OpenAI functions and the Node.js SDK. This browser-friendly tutorial uses simple tools like IP API and Open Meteo—no server required. Build it step-by-step and see AI take real actions!
Creating a Powerful Voice Assistant with OpenAI's Agents SDK
Dreaming of a unified voice assistant that handles every user query seamlessly? OpenAI's Responses API, Agents SDK, and Voice agents make it easier than ever. Follow this guide to build a triage-based voice assistant with search, knowledge, and account capabilities.
ElatoAI: Crafting Realtime Speech AI Agents for ESP32 with Arduino
Build your own AI voice agent device leveraging OpenAI's Realtime API, ESP32, secure WebSockets, and Deno Edge Functions for seamless global conversations lasting over 10 minutes. This comprehensive guide walks you through hardware assembly, software setup, and deployment for instant speech-to-speech interactions.
Hands-On Cookbook for Selecting and Deploying OpenAI Models in Real-World Scenarios
Discover practical frameworks and examples to pick the ideal OpenAI model—GPT-4.1, o3, or o4-mini—for your workloads. Tailored for engineers and practitioners, this guide delivers quick decisions, code snippets, and paths from prototype to production. Stay ahead with current capabilities and industry-focused use cases.
Building Voice Agents with MCP: The Agentic Framework Cookbook
Unlock the power of Model Context Protocol (MCP) to create scalable, voice-enabled agents for enterprise needs like insurance support. This guide delivers an end-to-end recipe using OpenAI Agents SDK for dynamic tool-calling and context management. Perfect for copilots and customer assistants across industries.
Crafting a Supply Chain Copilot with OpenAI Agent SDK and Databricks MCP Servers
Unlock real-time supply chain insights with an AI agent that tackles inventory checks, delay predictions, and revenue risks. This guide uses OpenAI Agent SDK and Databricks Managed MCP to query diverse data sources for proactive recommendations. Deploy a conversational interface that empowers stakeholders with instant answers.
Harnessing PLANS.md for Extended Multi-Hour Problem Solving
Unlock Codex's potential for tackling complex, time-intensive tasks with PLANS.md. This guide shows how to create living design documents that steer the model through research, design, and implementation for hours from one prompt. Set up AGENTS.md and dive into a proven PLANS.md template for success.
Revitalizing Legacy Codebases with OpenAI's Codex CLI
Unlock the power of OpenAI's Codex CLI to streamline codebase modernization. This guide uses a COBOL investment portfolio as an example, outlining a 5-phase approach with ExecPlans for clear, auditable, and scalable transformations. Let Codex manage tedious refactors while you oversee architecture and rules.
Transforming Legacy Codebases with OpenAI Codex CLI
Unlock the power of OpenAI's Codex CLI to modernize complex legacy systems effortlessly. This practical guide walks you through a phased approach using ExecPlans, agents, and precise prompts to ensure understandable, auditable, and repeatable code transformations for stacks like COBOL.
Harnessing PLANS.md for Multi-Hour Coding Challenges with Codex
Empower Codex and the recommended `gpt-5.2-codex` model to conquer complex, time-intensive tasks through detailed PLANS.md guides. These living design documents let you preview and direct long-running implementations, proven to sustain over seven hours from a single prompt. Integrate via AGENTS.md for reliable AI-driven project success.
Transforming Legacy Codebases with Codex: A Practical Guide
Unlock the power of OpenAI's Codex CLI to modernize complex legacy systems like COBOL portfolios. This structured cookbook uses ExecPlans and agents for clear, auditable, and repeatable refactoring, letting engineers focus on high-level design while Codex handles the details. Start with a pilot flow to ensure behavioral parity and scale effortlessly.
MCP Servers + WebSockets: Live Chatbots Powered by Claude
Supercharge Claude AI chatbots with MCP servers and WebSockets for true bidirectional, low-latency conversations. Stream responses in real-time without polling delays.
Claude API + LangGraph: Build Stateful AI Agents in Python
Discover how to build powerful, stateful AI agents using Claude API and LangGraph. This guide provides step-by-step Python code for persistent, multi-step automation that retains context across intera
Claude Sonnet 4 vs Llama 4: 2025 Benchmarks for Agentic Coding Tasks
Claude Sonnet 4 dominates Llama 4 in 2025 agentic coding benchmarks, nailing multi-step tasks with 92% success vs 78%. But costs and speed tell a different story—let's break it down.
Prompt Engineering: Claude's XML Tool Calling for Complex Multi-Step Tasks
Unlock Claude's power for complex tasks with XML-structured tool calling—parse precise multi-step actions effortlessly. Perfect for e-commerce automation and research agents.
No-Code Claude Agents in Bubble: Custom Workflows Without Coding
Build intelligent Claude AI agents in Bubble without coding. This tutorial shows how to create lead scoring and dynamic forms for smarter workflows.
Claude API + n8n: No-Code Multi-Agent Workflows for Marketing Campaigns
Unlock scalable marketing magic: build no-code multi-agent workflows in n8n with Claude API to automate content creation, A/B testing, and analytics – no dev skills needed.
MCP Servers Unleashed: Extend Claude with Custom Web3 and Database Tools
Supercharge Claude AI agents with custom MCP servers connecting to Web3 blockchains and real-time databases. This tutorial guides you through building, deploying, and integrating them for agentic work