Coding & Development Tools

This is our largest category, and it maps the fastest-moving part of the AI market. The tools cluster into six jobs. AI-first code editors — Cursor and Devin Desktop (formerly Windsurf) — rebuild the IDE around the model. Terminal agents — Claude Code, Gemini CLI, and Aider — work from the command line and handle multi-file tasks autonomously. Autocomplete assistants — GitHub Copilot, Codeium, Tabnine — predict your next lines inside the editor you already use. App generators — Bolt.new, Lovable, v0, and Replit — turn prompts into working prototypes, aimed as much at non-engineers as developers. Review and quality tools — CodeRabbit, Graphite, Greptile, Qodo — catch bugs in pull requests. And autonomous agents — Devin, Factory, Sweep — take tickets and attempt the whole task.

The dividing line that matters: assistance versus autonomy. Autocomplete makes you faster; agents try to do the job while you review. Most professional teams now run both — an agent for well-scoped tasks, autocomplete for everything else.

Pricing follows the split. Most tools are freemium with seat-based plans, while agent-heavy tools increasingly charge for usage on top. The open-source options — Aider, Cline, Roo Code, and Gemini CLI — are free; you pay only for the model API. When comparing, weight codebase context above everything: a tool that understands your repo beats a smarter model that doesn't.