Tools for AI Assisted Coding

Tool Categories

Comparison

Which Tool for Which Task?

Tool Details

GitHub Copilot

Lightweight autocomplete + chat that lives inside your existing IDE (VS Code, JetBrains, Neovim).

• Very fast inline suggestions — shines in TypeScript, Python, Go, Java, C#
• Deep GitHub integration — PR summaries, test generation, diff-aware context
• Low friction — install, log in, start coding
• Less focused on multi-file agent tasks than AI-first IDEs (but improving fast)

Project rules: .github/copilot-instructions.md

Cursor

AI-first IDE (VS Code fork) built around agents and codebase-level understanding.

• Project-wide edits — great at refactors and keeping style consistent across files
• Agent mode — multi-step tasks: edit files, run commands, fix lints/tests
• Multi-model support — plug in OpenAI, Anthropic, Google, pick per task
• Semantic search + project context give it strong “feel” for your code
• Works best if you invest time into project configuration

Project rules: .cursorrules

Claude Code

Autonomous coding agent (CLI) that reads large codebases and executes multi-step plans.

• Long-context reasoning — strong at architecture reviews, migration plans, non-trivial refactors
• Integrations — CLI, VS Code, JetBrains, desktop app
• Plan mode — design first, execute after approval
• Best paired with an IDE rather than used as your only tool

Project rules: CLAUDE.md, AGENTS.md

Windsurf (Codeium)

AI-powered editor with aggressive autocomplete and agent features.

• Supercomplete — multi-line completions with Tab-to-Jump and Tab-to-Import
• Cascade agent — runs workflows, searches web/docs, automates repetitive tasks
• Broad ecosystem — VS Code, JetBrains, Visual Studio, Vim, Jupyter
• Feature set evolves quickly; expect to experiment

ChatGPT

General-purpose assistant that’s also strong at code.

• Great for learning, explanations, and step-by-step walkthroughs
• Code Interpreter — run Python in chat for data analysis and scripts
• Canvas — reason over multiple files in a structured workspace
• Custom GPTs — build project-specific or domain-specific helpers
• Not tied to your IDE — pair it with a coding tool for day-to-day work

Codex (OpenAI)

Open-source CLI agent similar to Claude Code, using OpenAI models.

• Full control — open-source, can self-host and customize
• Agent capabilities — multi-step tasks, refactors, code analysis
• IDE integrations — VS Code, JetBrains, or standalone CLI
• Requires more setup than turnkey solutions

v0 (Vercel)

UI-first code generator for React / Next.js frontends.

• Design-to-code — prompts or screenshots → React components with Tailwind + shadcn/ui
• Instant browser previews — edit and see changes live
• One-click Vercel deploy
• Frontend focused — not for backend or infrastructure

Lovable

Browser-based AI app builder for full-stack applications.

• Chat-to-app — describe your vision or upload screenshots → working prototype
• Full-stack in browser — frontend, backend, deployment without leaving the platform
• Template library — portfolios, e-commerce, blogs, event platforms
• Best for smaller-scale projects — export to a proper repo for heavy customization

Continue (Open Source)

Open-source AI extension for VS Code and JetBrains.

• Bring-your-own-model — OpenAI, Anthropic, self-hosted (Llama, Mistral), on-prem
• Maximum control over models, privacy, and data processing
• In-editor chat with context-aware code understanding
• Power depends on which underlying model you connect

JetBrains AI Assistant

Native AI layer built into JetBrains IDEs.

• Deep IDE awareness — understands refactorings, inspections, project structure
• Inline completions, code reviews, context-aware chat
• Strong for JVM / enterprise stacks (IntelliJ, Rider, GoLand)
• Makes most sense if JetBrains is already your primary IDE

Getting Started

For this workshop, we recommend starting with one of these combinations:

Pick one, get comfortable with it, then explore others as needed.