This topic covers the evolving ecosystem of AI coding agents and developer terminal assistants — tools that generate, complete, explain, and act on code inside IDEs, web editors, and agent pipelines. As of 2026‑05‑29 the landscape emphasizes two converging trends: in‑IDE copilots that augment developer workflows (e.g., GitHub Copilot, JetBrains AI Assistant, Replit Assistant) and agentic platforms that chain LLMs, state, and tool calls to perform multi‑step tasks (e.g., LangChain/LangGraph, GPTConsole, Windsurf/Cascade agents). Key tool categories include: IDE‑integrated copilots for context‑aware completions and refactorings; open‑source and research code models for fine‑tuning and on‑prem inference (Salesforce CodeT5, CodeGeeX, WizardCoder families); AI‑native web IDEs that combine hosting, previews and assistants (Replit, Windsurf); and developer platforms for building, testing, and deploying agents with memory and lifecycle controls (LangChain, GPTConsole). Peripheral tooling such as analytics and prompt‑performance measurement (Peec AI) is increasingly part of productionization. Relevance: organizations now evaluate tradeoffs between proprietary hosted models (Codex, Claude Code, Grok Build) and open models for costs, licensing, security, and customization. Production adoption requires not just model quality but agent orchestration, evaluation, reproducibility, and CI/CD integration. Practical concerns — vulnerability scanning, licensing of training data, latency, and multi‑model routing — factor into tool choice. This topic synthesizes current capabilities and deployment patterns so teams can compare copilots, agent frameworks, and code LLMs by function: local customization, in‑editor productivity, agentic automation, and production readiness.